Do Prices Determine Vertical Integration?
Evidence from Trade Policy∗
Laura Alfaro
Harvard Business School and NBER
Paola Conconi
Université Libre de Bruxelles (ECARES) and CEPR
Harald Fadinger
University of Vienna
Andrew F. Newman
Boston University and CEPR
May 2013
Abstract
What is the relationship between product prices and vertical integration? While the literature has focused on how integration affects prices, this paper shows that prices can
affect integration. Many theories in organizational economics and industrial organization
posit that integration, while costly, increases productivity. If true, it follows from firms’
maximizing behavior that higher prices cause firms to choose more integration. The reason
is that at low prices, increases in revenue resulting from enhanced productivity are too
small to justify the cost, whereas at higher prices, the revenue benefit exceeds the cost.
Trade policy provides a source of exogenous price variation to assess the validity of this
prediction: higher tariffs should lead to higher prices and therefore to more integration. We
construct firm-level indices of vertical integration for a large set of countries and industries
and exploit cross-section and time-series variation in import tariffs to examine their impact
on firm boundaries. Our empirical results provide strong support for the view that output
prices are a key determinant of vertical integration.
JEL classifications: D2, L2,
Keywords: Theory of the firm, vertical integration, product prices.
∗
Some of the material in this paper appeared in “Trade Policy and Firm Boundaries,” IED DP 199, which it
supersedes. We are grateful to Daron Acemoglu, Robby Akerlof, Nick Bloom, Bob Gibbons, Davin Chor, Gene
Grossman, Maria Guadalupe, Ali Hortaçsu, Louis Kaplow, Patrick Legros, Margaret Mcmillan, Emanuel Ornelas,
Raffaella Sadun, Chad Syverson, John Van Reenen, Chris Woodruff and numerous seminar and conference
participants for their valuable suggestions and comments. Research funding from the FNRS and the European
Commission (PEGGED collaborative project) is gratefully acknowledged by Conconi. For their hospitality,
Newman is grateful to the Yale Economic Growth Center and the Brown Population Studies and Training
Center. We thank Francisco Pino, Andrea Colombo and Qiang Wang for excellent research assistance.
1
Introduction
The relationship between vertical integration and product prices has generated much debate
among economists and policy makers. Two strands of thought, broadly opposed, have emerged.
The foreclosure view emphasizes the possibility of reduced competition, with accompanying increases in product prices.1 The efficiency view, by contrast, maintains that integration increases
productivity, thereby reducing prices.2 Discussion usually revolves around which of these opposing effects is likely to dominate in a particular market or merger case. Either way, causality
runs from vertical integration to prices.
Efficiency theories have another ramification, however, which can predict a positive association between prices and integration. This implication is easiest to see if one assumes perfect
competition in the product market: supposing that integration increases productivity, a pricetaking firm will integrate more when the price of its product is higher. At a low price, the
increment in revenue resulting from integration is too small to justify any (fixed) cost integration may have; at a high price, the incremental revenue is large enough to make integration
worthwhile. More generally, if integration affects efficiency, there will be a force running in the
opposite direction, from prices to vertical integration.3
In this paper, we provide evidence that this mechanism is operative in a wide range of
industries around the world. The main empirical challenge is to find sources of price variation
that are exogenous to firms’ vertical integration decisions. Our strategy is to exploit crosssection and time-series variation in Most-Favored-Nation (MFN) tariffs applied by GATT/WTO
members. Since tariffs raise product prices in the domestic market, they should lead to more
vertical integration among firms selling in that market.
MFN tariffs can be taken as exogenous to vertical integration for several reasons. First,
they emerge from long rounds of multilateral trade negotiations. At the end of each round,
GATT members commit not to exceed certain tariff rates (tariff bounds); if a country raises its
tariffs above the bound level, other countries can take it to dispute settlement. Second, they
1
Key theoretical contributions on market foreclosure include Salinger (1988), Ordover, Saloner, and Salop
(1990), Hart and Tirole (1990) and Bolton and Whinston (1993). Market foreclosure concerns have been enshrined in anti-trust policies and have motivated policies such as “divorcement” legislation. See, for example, the
guidelines on the assessment of non-horizontal mergers in the United States (1984 Merger Guidelines) and in the
European Union (Council Regulation 2008/C 265/07)
2
Numerous channels have been identified whereby integration enhances productivity. Technological synergies and efficiencies in asset use are frequently cited by policy makers and antitrust defendants. Organization
economists have emphasized other benefits, and often associated costs: reductions in the costs of transactions,
adaptation, or opportunism (Williamson, 1971, 1975; Klein, Crawford, and Alchian, 1978); better multitasking
incentives (Holmström and Milgrom, 1991); alignment of control and incentives (Grossman and Hart, 1986; Hart
and Moore, 1990); or improved coordination (Hart and Holmström, 2010). A complementary class of theories
emphasize allocative, rather than productive, efficiency gains achieved by the elimination of double markups,
though these shall not concern us. See Perry (1989) and Riordan (2008) for further discussion.
3
Legros and Newman (2013) seems to be first to point out this relationship, in the context of a “propertyrights” model of integration choices by competitive firms with or without entry. As discussed in Section 3, it can
also be extended to imperfect competition.
1
must be applied in a non-discriminatory manner to imports from all countries, which severely
limits negotiators’ flexibility to respond to political pressure.4 As a result, governments resort to
administrative measures for the regulation of imports, such as anti-dumping and countervailing
duties, to respond to short-term political pressure (e.g. Finger, Hall and Nelson, 1982). Third,
they are persistent, significantly more so than integration choices. In our main analysis, we
study vertical integration of firms in 2004. In that year, the prevailing tariffs resulted from the
eight-year Uruguay Round of trade negotiation that was completed ten years earlier.5 Finally,
while firm size and industry concentration might lead to higher final good tariffs by alleviating
free-riding problems in lobbying (Mitra, 1999; Bombardini, 2008), there is little reason to believe
that vertical structure should have such an effect.6
The effect of product prices on integration that we propose should apply broadly, to many
different markets. We therefore draw our evidence from the WorldBase dataset of Dun and
Bradstreet (D&B), which contains information on firms in many different countries and industries. This approach allows us to exploit cross-country and cross-sector variation in MFN tariffs.7
WorldBase contains listed and unlisted plant-level observations for a large set of countries and
territories. For each plant, the dataset includes information about its production activities (at
the 4-digit SIC level) and ownership (e.g. domestic or global parent). To measure vertical
integration, we apply the approach of Fan and Lang (2000): combining information on firms’
production activities with input-output tables, we construct firm-level vertical integration indices, which measure the fraction of inputs used in the production of a firm’s final good that
can be produced in house.
Our empirical results provide strong support for the view that output prices are a key determinant of vertical integration. We find that the higher is the MFN tariff applied by a country on
the imports of a given product, the more vertically integrated are firms producing that product
in that country. The effect is larger when we would expect organizational decisions to be more
responsive to import tariffs (i.e. for firms that only serve the domestic market) and in sectors in
which MFN tariffs have a larger impact on domestic prices.8
4
The MFN treatment obligation stipulated in Article I of the General Agreement on Tariffs and Trade (GATT)
forbids members to discriminate between trading partners. It requires that equal treatment be afforded to all
imported goods, irrespective of their origin.
5
By 2004, most GATT/WTO members had reduced their tariffs to meet the binding obligations agreed to in
1994, at the end of the Uruguay Round of trade negotiations (Bchir, Jean and Laborde, 2006).
6
If anything, vertical integration will tend to make free-riding problems in lobbying for protection more severe:
vertically integrated suppliers will have a diversity of interests that would weaken their lobbying incentives;
moreover, coordination to lobby will be harder among suppliers that are in the same sector but belong to
different firms than among independent suppliers. Our results are unaffected when controlling for firm size and
industry concentration.
7
The GATT non-discrimination principle implies that there is only one MFN tariff rate per industry in each
country; the length of multilateral trade rounds — and the long gaps between them — imply that MFN tariffs
vary little over time. Most of the variation is thus within countries across industries and within industries across
countries (see Section 4.4).
8
In our main empirical analysis, we focus on firms that have plants only in one country. There are three
main reasons for this choice. First, these firms provide a cleaner analysis of the effects of product prices on
2
Our estimates imply that price changes can have large effects on firm boundaries. Depending
on specification, we obtain estimates of the tariff elasticity of vertical integration in the range
0.02 up to 0.1. Though this may seem inconsequential at first blush, given that tariffs are
expressed in ad-valorem terms rather than in units, and that the mean tariff is on the order of
5 percent, these translate into much larger price elasticities, in the range of 0.4 - 2.9
In our theory, in which firms are price takers, import tariffs affect firms’ organization through
their effect on domestic prices. However, tariffs can also have an impact on the degree of
competition faced by domestic firms, which may also shape vertical integration decisions (Aghion,
Griffith, and Howitt, 2006). To isolate the effect of product prices, we restrict our analysis
to highly competitive sectors, in which tariffs will have little or no effect on the degree of
competition, obtaining even stronger results.
Another possible explanation for the positive effect of tariffs on vertical integration is that,
in the presence of credit constraints, protected firms may have more disposable cash to acquire
their suppliers. This mechanism would be expected to be strongest where credit markets are
least efficient, or in industries which are most financially dependent. We verify that the effect of
tariffs on integration does not vary with either of these factors, as captured by standard measures
of financial development and financial dependence.
To establish a causal link between tariffs and vertical integration, we also show that our results
are not driven by omitted variables, which could be correlated with both vertical integration
decisions and MFN tariffs on final products (e.g. the degree of industry concentration, input
tariffs, and tariffs in export markets). Our results are unaffected when including these controls
in our analysis. They are also robust to constructing vertical integration indices in different
ways, using alternative econometric methodologies, and focusing on different samples of firms
and countries.
An alternative strategy to verify the impact of product prices on firm boundaries is to exploit
time variation in import tariffs, examining the effects of trade liberalization reforms — major
unilateral or multilateral liberalization episodes, or the creation of regional trade agreements —
on vertical integration decisions. The challenge with implementing this strategy is data availability, since we can only construct firm-level vertical integration measures for recent years, during
which there have been few trade liberalization reforms.10 The only major trade liberalization
firms’ ownership structure; in the case of firms operating in many markets, it is harder to identify the relevant
prices and tariffs. Second, focusing on national firms avoids issues having to do with the strategic behavior of
multinationals across markets (e.g. transfer pricing, tariff jumping, export platforms). Finally, when integration
occurs across international borders as opposed to within them, trade policy can alter bargaining power (surplus
division) among suppliers as well as the value of what they jointly produce, further complicating the predicted
effects (Ornelas and Turner, 2008; Antras and Staiger, 2012).
9
Another way to get a sense of the effects of prices on organization would be to instrument prices with MFN
tariffs. However, this would require comparable cross country data on domestic prices, which are very difficult
to obtain (see Bradford, 2003). Moreover, even if these data were available, it would be hard to argue that the
exclusion restriction is satisfied. The reduced-form approach followed in this paper avoids both problems.
10
The first year for which we can use WorldBase to construct vertical integration indices is 1999. Important
3
episode that has occurred in recent years is arguably the entry of China into the WTO in 2001: to
be accepted as a WTO member, China agreed to undertake substantial reductions in its import
tariffs. We examine the organizational effects of these tariff changes, comparing the ownership
structure of Chinese firms before and after WTO accession (in 1999 and 2007). Consistent with
the predictions of our theoretical model, we find that firm-level vertical integration has fallen
more in sectors that have experienced larger tariff cuts.
Finally, we also investigate the effect of trade policy on the degree of organizational convergence across countries. Our theory suggests that countries with similar domestic price levels
should have firms with similar ownership structures. In line with this prediction, we show that
differences in vertical integration across countries are significantly larger in sectors in which
differences in MFN tariffs (and therefore differences in domestic prices) are larger. Moreover,
we find that differences in vertical integration indices are smaller for country pairs engaged in
regional trade agreements.11 This effect is stronger for customs unions, which impose common
external tariffs vis-à-vis non-members and should therefore achieve greater price convergence.
This paper focuses on vertical integration, which involves complementary goods linked in
a buyer-supplier relationship. In principle, the theoretical mechanism we investigate may also
apply to horizontal integration, which involves substitute goods, or lateral integration, involving
goods sold in separate markets that are complementary either in production or consumption.
To the extent that these forms of integration also are costly but enhance productive efficiency,
we should expect firms to be more integrated in these other dimensions when tariffs — and thus
product prices — are higher. Data limitations, as well as lack of an unconfounded measure of
horizontal integration make it difficult to apply the methodology to these other cases at present,
as discussed in Section 4.3. Thus we feel that vertical integration provides the cleanest test of
the theory.
The rest of the paper is organized as follows. Section 2 discusses the related literature.
Section 3 presents a theoretical framework to guide our empirical analysis. Section 4 describes
our data. Section 5 and 6 present our main results, exploiting cross-sectional and time series
variation in tariffs. Section 7 analyzes the impact of trade policy on the degree of cross-country
organizational convergence. The last section concludes.
multilateral or regional trade liberalization episodes, such as the Uruguay Round Agreements, the NAFTA free
trade area between the U.S., Canada and Mexico, or the MERCOSUR customs union between Argentina, Brazil,
Paraguay and Uruguay, all occurred in the early or mid-nineties.
11
Under Article I of the GATT, countries have to apply the same MFN tariff to all trading partners. Preferential
treatment can only be granted to partners in regional trade agreements (under Article XXIV of the GATT) or
developing countries (in the context of the Generalized System of Preferences, under the Enabling Clause).
4
2
Related literature
Understanding vertical integration decisions has been a fundamental concern of organization
economics since Coase (1937)’s seminal paper. We have already mentioned (footnote 2) some of
the seminal contributions, both formal and informal, that have shaped economists’ understanding
of how ownership structure affects productivity of individual firms. Recent theoretical work has
embedded models of firms into market settings to study how firms’ boundary choices are affected
by market conditions. In particular, market thickness, demand elasticities, and terms of trade
in supplier markets may have an impact on firms’ vertical integration decisions (e.g. McLaren,
2000; Grossman and Helpman, 2002; Legros and Newman, 2008, 2013). So far, evidence on the
implications of these models is sparse. This paper shows that market conditions — in particular,
the level of product prices — do affect vertical integration decisions.
There is a very large empirical literature that examines the determinants of firms’ vertical
integration decisions (i.e. firm boundaries/ownership structure), usually with a view to assessing
the importance of different tradeoffs that determine firm boundaries, or to examining effects of
vertical integration on market outcomes (for an excellent survey, see Lafontaine and Slade,
2007). Most studies focus on single industries.12 In this literature, Hortaçsu and Syverson
(2007) concentrate on the U.S. cement industry and examine whether vertical integration leads
to higher prices. In contrast with the predictions of market foreclosure theories, they find that
more integration leads to lower prices; they do not address with the opposite direction of causality
that is our concern.
A few studies examine a single country. For example, Acemoglu, Aghion, Griffith and Zilibotti
(2010) use data on British manufacturing plants to study the relationship between vertical
integration and rates of innovation. Aghion, Griffith and Howitt (2006) investigate whether the
propensity for firms to vertically integrate varies systematically with the extent of competition
in the product market.
As for multi-country studies, one stream of the literature has analyzed other aspects of organization, such as management practices or the degree of delegation within firms. Bloom and
Van Reenen (2007) study managerial practices in medium-sized manufacturing firms in the US
and Europe (France, Germany and the UK), finding that best practices are strongly associated
with superior firm performance. Bloom, Sadun and Van Reenen (2010), using survey data on
manufacturing firms across a dozen countries, reveal that greater product market competition
increases decentralization. Bloom, Sadun and Van Reenen (2012) using survey data they collected from several countries to show that firms headquartered in high trust regions are more
likely to decentralize. Guadalupe and Wulf (2012) show that the 1989 Canada-United States
12
These include the seminal papers by Stuckey (1983) on integration between aluminum refineries and bauxite
mines and Joskow (1987) on ownership arrangements in electricity generating plants, as well as the more recent
studies by Baker and Hubbard (2003, 2004) on the trucking industry, Woodruff (2002) on Mexican footwear; or
Forbes and Lederman (2009, 2011) on airlines.
5
Free Trade Agreement (CUSFTA) led large U.S. firms to flatten their hierarchies. Other papers
have studied how trade liberalization, by increasing the degree of competition, affects the number
of horizontally differentiated product varieties a firm chooses to manufacture (Eckel and Neary,
2010; Bernard, Redding and Schott, 2011).
Various papers examine whether goods are sold within or across firm boundaries in the
global economy (e.g. Antras, 2003; Nunn, 2007, Diez, 2010). This literature considers the
organizational choices of multinational firms and highlights the role of contract enforcement and
relationship-specific investments. By contrast, we focus on the organizational choices of firms
that operate in a single country.
In terms of data and methodology, our analysis is closely related to the paper by Acemoglu, Johnson and Mitton (2009), who study the determinants of vertical integration using a
cross-section of D&B data for 93 countries, emphasizing the role of financial development and
contracting costs. Ours is the first paper to investigate the impact of product prices on vertical
integration decisions.
3
Theoretical Framework
The fundamental logic of how the price level influences the degree of integration is most simply
illustrated with a “reduced form” model in which vertical integration has three main features:
(i) it enhances productivity; (ii) it does so at a cost; (iii) the cost is independent of product price.
The first assumption is the defining attribute of efficiency theories of vertical integration. The
second is necessary if there is anything to discuss: without it, given the first assumption, firms
would always integrate to the maximal extent. The third is commonly made, either directly, or
derived from more fundamental assumptions.
Consider a price-taking enterprise that requires N ≥ 2 inputs to produce a final good priced
at P . Before production, the enterprise chooses n, the number of inputs that will be produced
inside a single firm.The effects of integration on productivity are modelled in the simplest possible
way: the cost of producing Q units of output is ψ(n)C(Q), where ψ(·) is decreasing and C(Q)
is increasing and convex. Thus, the more integrated the firm (higher is n), the lower the unit
cost of producing Q.
Integration is costly (if not, we would always have n = N ). Let Φ(n) be the cost of integrating n units into the firm Φ(·) is increasing and Φ(0) = 0. This function captures different
types of costs (e.g. legal, bureaucratic, monitoring; or private costs of effort, subordination, or
conformity). Note that here the integration cost is independent of P and Q. The enterprise’s
net profit is then
P Q − ψ(n)C(Q) − Φ(n).
(1)
6
The enterprise chooses n and Q to maximize profit, taking P as given.
Since ψ(·) is decreasing and C(·) is increasing, the profit has (strictly) increasing differences (is
strictly supermodular) in the choice variables n and Q. Basic principles of monotone comparative
statics (e.g. Topkis, 1998; Vives, 2000) tell us that optimal choices of these variables will therefore
co-vary. Since profit has (strictly) increasing differences in P and Q, the optimal quantity Q,
and therefore the optimal degree of integration n, will increase with P . The intuition is that a
given productivity increase via integration is more valuable when the price of output is higher,
so higher prices increase integration incentives.13
In our empirical analysis, we will exploit variation in import tariffs to assess the validity
of this prediction. We expect a positive relationship between tariffs and vertical integration:
given a world price p, an ad-valorem tariff t raises the domestic price for a small country to
P = (1 + t)p. Higher tariffs increase the gains from integration for firms located in the domestic
market. Of course, in practice, firms will experience changes in other variables of the model,
such as demand (which can affect P without affecting t), technology (C(Q)), or the integration
cost Φ(n) at frequencies much higher than changes in MFN tariffs; indeed it is the relative
infrequency with which MFN tariffs change that makes them a good proxy for the domestic
price.
The basic prediction does not hinge on the assumption of perfect competition. The same logic
applies if the domestic firm is a monopolist: as long as the the tariff-augmented price is below
the monopoly price, an increase in the tariff will increase the monopolist’s revenue, inducing
more integration.14
Several corollaries and qualifications follow from this basic logic. First, the impact of import
tariffs on integration choices should be stronger for firms that only serve the domestic market,
since their profit depends only on the domestic price; the effect should be weaker for firms that
also serve foreign markers (exporting firms and multinationals), since their profit and integration
decisions depend partly on prices prevailing in other countries.
Second, the impact of trade policy on the degree of vertical integration should also depend on
the extent to which import tariffs affect domestic prices. In particular, the higher is the share of
imports that are subject to the tariff (i.e. the lower is the share of goods imported duty-free from
regional trading partners), the larger should be the effect of tariffs on prices and organization.
13
To be sure, in some models, particularly those in which incentives play a role, the extent of the efficiency
gains may depend on other variables besides n, such as the price P or the distribution of the profits among
the various production units (Legros and Newman, 2013). These more general specifications may also lead to
nonlinearities and non-monotonicities in the predicted relationship between integration and price. However, these
complications do not affect the basic contention that product prices influence integration decisions. We did not
find systematic evidence of these more complex patterns in our data.
14
In case of monopoly, the firm faces an inverse demand P (Q) rather than a constant price P , but cannot
charge more than the tariff-augmented world price p(1 + t). It will charge less if the monopoly price P (Q) is
less than p(1 + t). Thus its objective is to maximize min {p(1 + t), P (Q)}Q − ψ(n)C(Q) − Φ(n), which has the
same properties (nondecreasing differences in t and Q, supermodularity in n and Q) as the competitive firm’s
objective.
7
Finally, the law of one price implies a law of one organization: trade policy should affect the
degree of organizational convergence across countries. In particular, vertical integration choices
should be more alike among countries in sectors in which their import tariffs (and thus their
domestic prices) are close. We also expect a tendency for convergence in prices and organizational
choices among members of regional trade agreements; this result should be stronger for customs
unions — in which members adopt common external tariffs – than in free trade areas — in which
differences in external tariffs and rules of origins reduce the extent of price convergence.15
To summarize, for the purpose of our empirical analysis, the main predictions of our theoretical framework are:
1. Higher import tariffs on final goods should induce domestic firms producing these goods
to be more vertically integrated.
2. The effect of tariffs on integration should be larger for firms serving only the domestic
market.
3. The effect of tariffs on integration should be larger in sectors in which a smaller fraction
of imports are exempt from the tariff.
4. Country pairs should have similar ownership structures in sectors where they face similar
levels of protection; regional trade agreements, especially customs unions, should display
similar ownership structures among members.
4
Dataset and variables
In Sections 5-7, we will provide evidence that product prices affect vertical integration decisions
in a wide range of countries and industries, in line with the above predictions. Focusing on
many countries and industries allows us to exploit MFN tariffs as a source of exogenous price
variation.16 In this section, we describe our dataset and the variables used in our empirical
analysis.
4.1
The WorldBase database
Increasingly, researchers use multi-country firm-level data to study issues of organization economics (e.g. Bloom and Van Reenen, 2007; Bloom, Sadun and Van Reenen, 2012). However,
cross-country empirical investigations at the firm level are notoriously challenging due to both
15
See Cadot, de Melo, and Olarreaga (1999), for a comparison of different types of regional trade agreements
and a discussion of rules of origin in free trade areas.
16
As discussed in the introduction, MFN tariffs are very persistent, i.e. vary little in between rounds of
multilateral trade negotiations. At a given point in time, given the GATT principle of non-discrimination (see
footnote 4), MFN tariffs vary across countries (for a given industry) and across industries (for a given country).
8
the lack of data and the difficulty of comparing the few high quality time-series datasets that
are available (mostly in rich countries). The reason for the data constraint is simple: economic
censuses of firms are infrequently collected due to high costs and institutional restrictions, especially in poor countries. No institution has the capacity or resources to collect census data for a
wide range of countries and periods. This is why researchers have to use other sources, such as
business “compilations” (registries, tax sources) or surveys.
To measure vertical integration, we use data from Dun & Bradstreet’s WorldBase, a database
covering public and private companies in more than 200 countries and territories.17 The unit of
observation is the establishment/plant. With a full sample, plants belonging to the same firm
can be linked via information on domestic and global parents using the DUNS numbers.18
The WorldBase dataset has been used extensively in the literature. Early examples include
Caves’ (1975) analysis of size and diversification patterns between Canadian and U.S. plants.
More recent uses include Harrison, Love, and McMillian (2004), Black and Strahan (2002),
Alfaro and Charlton (2009), and Acemoglu, Johnson and Mitton (2009). One of the advantages of
WorldBase compared to other international datasets is that it is compiled from a large number of
sources (e.g. partner firms, telephone directory records, websites, self-registration). Admittedly,
sample coverage may vary across countries, but this problem can be mitigated by focusing on
manufacturing firms above a size threshold of twenty employees (see discussion below).19
4.2
The sample
Our main sample is based on the 2004 WorldBase dataset (for the analysis of China’s accession
to the WTO, we use data from 1999 and 2007). As mentioned above, the unit of observation in
WorldBase is the establishment/plant, a single physical location at which business is conducted
or services or industrial operations are performed.
For each establishment, we use different categories of data recorded in WorldBase:
17
WorldBase is the core database with which D&B populates its commercial data products, including Who Owns WhomTM , Risk Management SolutionsTM , Sales & Marketing SolutionsTM , and Supply
Management SolutionsTM . These products provide information about the “activities, decision makers, finances, operations and markets” of the clients’ potential customers, competitors and suppliers.The dataset
is not publicly available but was released to us by Dun and Bradstreet. For more information see:
http://www.dnb.com/us/about/db database/dnbinfoquality.html.
18
D&B uses the United States Government Department of Commerce, Office of Management and Budget,
Standard Industrial Classification Manual 1987 edition to classify business establishments. The Data Universal
Numbering System — the D&B DUNS Number — introduced in 1963 to identify businesses numerically for dataprocessing purposes, supports the linking of plants and firms across countries and tracking of plants’ histories
including name changes.
19
Other datasets use different methodologies in different countries. For example, the Amadeus dataset, provided
like Orbis by Bureau Van Dijk, uses data from the national public body in charge of collecting the annual accounts
in some countries (e.g. the UK) and collects it directly from firms in other countries (most of Eastern Europe).
Because of different disclosure requirements, the amount and type of information also varies among countries.
See Alfaro and Charlton (2009) for a more detailed discussion of the WorldBase data and comparisons with other
data sources.
9
1. Industry information: the 4-digit SIC code of the primary industry in which each establishment operates, and for most countries, the SIC codes of as many as five secondary
industries, listed in descending order of importance.
2. Ownership information: information about the firms’ family members (number of family
members, domestic parent and global parent).20
3. Location information: country, state, city, and street address of each family member (used
to link establishments within a family to the relevant tariff data).
4. Basic operational information: sales and employment.
5. Information on the trade status (exporting/non-exporting).
We carry out the analysis at the firm level, using DUNS numbers to link plants that have the
same ultimate owner. As discussed below, however, since the overwhelming majority of firms in
our sample have only one establishment, the qualitative results of our analysis are unaffected if
we measure vertical integration at the plant level or restrict the analysis to single-plant firms.
We exclude countries and territories with fewer than 80 observations. We further restrict the
sample to Word Trade Organization (WTO) members for which we have data on tariffs/regional
trading arrangements (see discussion below). Table A-1 in the Appendix lists the countries
included in our main sample.21 In robustness checks, we restrict the analysis to two subsamples
of countries: members of the OECD, and countries for which we have information on at least
1000 plants.
We focus on manufacturing firms (i.e. firms with a primary SIC code between 2000 and 3999),
which best fit our theory of vertical integration and for which tariff data are widely available. We
exclude firms that do not report their primary activity, government/public sector firms, firms
in the service sector (for which we have no tariff data) or agriculture (due to the existence of
many non-tariff barriers), and firms producing primary commodities (i.e. mining and oil and
gas extraction).
We further exclude firms with less than 20 employees, as our theory is less apt to apply to selfemployment or small firms with little prospect of vertical integration (see also Acemoglu, Aghion,
Griffith and Zilibotti, 2010). Restricting the analysis to firms with more than 20 employees also
enables us to correct for possible differences in the collection of data on small firms across
countries (see Klapper, Laeven, and Rajan, 2006).
In our main sample, we focus on firms that are located only in one country. This provides
a cleaner setting to verify the predictions of our theoretical model, since the degree of vertical
20
D&B also provides information about the firm’s status (joint-venture, corporation, partnership) and its
position in the hierarchy (branch, division, headquarters).
21
Further restrictions were imposed by data availability constraints related to the control variables, as explained
in the next subsections.
10
integration of these firms should depend primarily on the price at which they sell their product
in their country. In the case of multinational corporations, on the other hand, it is harder to
identify the relevant prices and tariffs. Moreover, focusing on national firms avoids issues having
to do with the strategic behavior of multinationals across markets (e.g. transfer pricing, tariff
jumping). Multinational corporations are included in the robustness analysis (see Section 5.3).
In order to link their organizational structure to domestic tariffs, we split them in separate
entities — one for each country — and use the primary activity of the respective domestic
ultimate to identify the relevant tariff.
We next describe the construction of the vertical integration indices and the other variables
used in our empirical analysis. Appendix Table A-2 presents summary statistics for all variables.
4.3
Vertical integration indices
Constructing measures of vertical integration is highly demanding in terms of data, requiring
firm-level information on sales and purchases of inputs by various subsidiaries of a firm. Such
data are generally not directly available and, to the best of our knowledge, there is no source for
such data for a wide sample of countries.
To measure the extent of vertical integration for a given firm, we build on the methodology
developed by Fan and Lang (2000). We combine information on plant activities and ownership
structure from WorldBase with input-output data to determine related industries and construct
the vertical integration coefficients Vjf,k,c in activity j, where k is the primary sector in which
firm f in country c is active.22
Given the difficulty of finding input-output matrices for all the countries in our dataset, we
follow Acemoglu, Johnson and Mitton (2009) in using the U.S. input-output tables to provide
a standardized measure of input requirements for each sector. As the authors note, the U.S.
input-output tables should be informative about input flows across industries to the extent that
these are determined by technology.23
The input-output data are from the Bureau of Economic Analysis (BEA), Benchmark IO
Tables, which include the make table, use table, and direct and total requirements coefficients
tables. We use the Use of Commodities by Industries after Redefinitions 1992 (Producers’ Prices)
tables. While the BEA employs six-digit input-output industry codes, WorldBase uses the SIC
22
In Acemoglu, Johnson and Mitton (2009), the sample is restricted to a maximum of the 30,000 largest records
per country in the 2002 WorldBase file (a limit imposed by cost constraints). For countries with more than 30,000
observations, they select the 30,000 largest, ranked by annual sales. Having information on the full sample of
establishments in WorldBase, we are able to link establishments to firms (see discussion below).
23
Note that the assumption that the U.S. IO structure carries over to other countries can potentially bias our
empirical analysis against finding a significant relationship between vertical integration and prices by introducing
measurement error in the dependent variable of our regressions. In addition, using the US input-output tables
to construct vertical integration indices for other countries mitigates the possibility that the IO structure and
control variables are endogenous. In robustness checks, we verify that our results are unaffected when restricting
the analysis to OECD countries, which are closer to the U.S. in terms of technology (See Section 5.3).
11
industry classification. The BEA website provides a concordance guide, but it is not a one-toone key.24 For codes for which the match was not one-to-one, we randomized between possible
matches in order not to overstate vertical linkages. The multiple matching problem, however, is
not particularly relevant when looking at plants operating only in the manufacturing sector (for
which the key is almost one-to-one).
For every pair of industries, i, j, the input-output accounts support calculation of the dollar
value of i required to produce a dollar’s worth of j. We construct the input-output coefficients
f
for each firm f , IOij
by combining the SIC information for each plant in each firm, the matching
f
codes, and the U.S. input-output information. Here, IOij
≡ IOij ∗ Iijf , where IOij is the inputoutput coefficient for the sector pair ij, stating the cents of output of sector i required to produce
a dollar of j, and Iijf ∈ {0, 1} is an indicator variable that equals one if and only if firm f owns
plants in both sectors i and j. A firm that produces i as well as j will be assumed to supply itself
with all the i it needs to produce j; thus, the higher IOij for an i-producing plant owned by the
firm, the more integrated in the production of j the firm will be measured to be. Adding up the
f
input-output coefficients IOij
for all inputs i, gives the firm’s degree of vertical integration in j.
To illustrate the procedure, consider the following example from Acemoglu, Johnson and
Mitton (2009) of a Japanese establishment with, according to WorldBase, one primary activity, automobiles (59.0301), and two secondary activities, automotive stampings (41.0201) and
f
coefficients for this plant are:
miscellaneous plastic products (32.0400).25 The IOij
Output (j)
Autos Stampings Plastics
Autos
0.0043 0.0000
0.0000
Input (i) Stampings 0.0780 0.0017
0.0000
Plastics
0.0405 0.0024
0.0560
SUM
0.1228 0.0041
0.0560
The table is a restriction of the economy-wide IO table to the set of industries in which
f
this establishment is active (i.e. it contains all of the positive IOij
values). For example, the
IOij coefficient for stampings to autos is 0.078, indicating that 7.8 cents worth of automotive
stampings are required to produce a dollar’s worth of autos. Because this plant has the internal
capability to produce stampings, we assume it produces itself all the stampings it needs.26 The
f
bottom row shows the sum of the IOij
for each industry. For example, given that 12.3 cents
worth of the inputs required to make autos can be produced within this plant, we would say
24
This concordance is available upon request. The BEA matches its six-digit industry codes to 1987 U.S. SIC
codes http://www.bea.gov/industry/exe/ndn0017.exe.
25
There is no concern of right censoring in the number of reported activities: only 0.94 percent of establishments
with primary activity in a manufacturing sector report the maximum number of five secondary activities.
26
Many industries have positive IOij coefficients with themselves; for example, miscellaneous plastic products
are required to produce miscellaneous plastic products. Any firm that produces such a product will therefore be
measured as at least somewhat vertically integrated.
12
that the degree of vertical integration for this plant is 0.123.
For firm f in primary sector k located in country c, we define the integration index in activity
j as
j
Vf,k,c
=
X
f,k
IOij
,
(2)
i
the sum of the IO coefficients for each industry in which the firm is active. Our measure of
vertical integration is based on the firm’s primary activity:
j
Vf,k,c = Vf,k,c
, j = k.
(3)
In the case of multi-plant firms, we link the activities of all plants that report to the same
headquarters and consider the main activity of the headquarters as the primary sector.
The approach we follow to identify vertical integration infers a firm’s level of vertical integration from information about the goods it produces in each of its establishments and the
aggregate input-output relationship among those goods. The advantage of this method is that
one need not worry about the value of intra-firm activities being affected by transfer pricing.
Another advantage is that using I-O tables avoids the arbitrariness of classification schemes that
divide goods into “intermediate” and other categories (Hummels, Ishii, and Yi, 2001).
One might be concerned about measuring vertical integration at the firm level, in light of
recent studies that find little evidence of trade between plants of the same firm.27 However, this
concern does not apply to our analysis. This is because 96% of the firms in our sample have
only one plant and 87% of plants are not connected (see Table A-2). The qualitative results of
our analysis are thus unaffected if we measure vertical integration at the plant-level or restrict
the analysis to single-plant firms.
Summary statistics for firm-level vertical integration are presented in Appendix Table A2, while Table A-3 reports average vertical integration indices by sector (at the 2-digit SIC
level).28 Our main sample consists of 196,586 domestic manufacturing firms with at least 20
employees located in 80 countries. The histogram in Figure 1 reports the distribution of vertical
integration indices for all firms in our main sample. According to our measure, most firms
produce relatively few inputs in house: the median vertical integration index is around 0.044
and the mean is 0.063.29
27
Hortaçsu and Syverson (2009) find little evidence of commodity shipments across plants in US nonmultinational firms. Similarly, Ramondo, Rappoport and Ruhl (2012) find the bulk of intra-firm trade between
affiliate and the U.S. parent to be concentrated among a small number of large affiliates.
28
The descriptive statics for our vertical integration measure are similar to Acemoglu, Johnson and Mitton
(2009). They report a mean of 0.0487 and median of 0.0334 for their vertical integration index. For our main
sample, the primary sector vertical integration index has a mean of 0.0627 and a median of 0.0437 (see Table
A-2). The ordering of industries by degree of vertical integration in Table A-3 is also similar to that reported by
Acemoglu, Johnson and Mitton (2009).
29
It should be noted that this measure does not consider payments to capital and labor services and is thus
13
0
5
Density
10
15
20
Figure 1: Firm-level vertical integration index
med.
.1
.2
.3
Firm−level vertical integration index
.4
.5
As mentioned in the introduction, the mechanisms outlined in our model could apply to other
types of integration. In this paper, we focus on vertical integration, which we can measure using
information available in our dataset on the primary and secondary activities of each firm and
applying the methodology developed by Fan and Lang (2000). While it would be interesting to
examine whether whether higher tariffs, by raising product prices, also lead firms to be more
horizontally or laterally integrated, these cases present some difficulties. One virtue of the FanLang index is that it depends only on firm characteristics and technology, and is in particular
invariant to industry composition. Firm-level measures of horizontal integration (e.g. firm size
to mean size ratio, measured variously) or industry-level measures (e.g. Herfindahl indices),
which need not be good proxies for firm-level integration anyway, are not. This makes those
measures vulnerable to selection effects: for instance, to the extent that firms are heterogeneous
in productivity, a reduction in tariffs forces less productive firms to exit, shifting market share
towards more productive firms (Melitz, 2003). This may result in a negative relationship between
tariffs and industry concentration, which is not directly related to our mechanism. As for lateral
integration, constructing a firm-specific measure would require sales of each plant by product
line for narrowly defined industries, which we do not observe in our dataset.
4.4
Tariffs and other trade variables
Our main strategy to empirically assess the impact of market prices on ownership structure
is to use data on most-favored-nation (MFN) tariffs applied by GATT/WTO members. As
argued in the introduction, these tariffs offer a source of price variation that is exogenous to firm
always less than unity. Indeed, in the U.S. an industry pays on average around 56% of gross output to intermediates, the rest being value added. Thus, even a fully vertically integrated firm in a typical sector would have an
index of only 0.56.
14
boundaries.
We collect applied MFN tariffs at the 4-digit SIC level for all WTO members for which
this information is available. We restrict the set of countries to WTO members, which are
constrained under Article I of the GATT by the MFN principle of non-discrimination: each
country c applies the tariff Tariff k,c to all imports of final good k that originate in other WTO
member countries. Preferential treatment is only allowed for imports originating from RTA
members or from developing countries (see discussion below).
The source for MFN tariffs is the World Integrated Trade Solution (WITS) database, which
combines information from the UNCTAD TRAINS database (default data source) with the WTO
integrated database (alternative data source). In our main empirical analysis, we use applied
MFN tariffs for 2004.30 The original classification for tariff data is the harmonized system (HS)
6-digit classification. Tariffs are converted to the more aggregate SIC 4-digit level using internal
conversion tables of WITS. Here, SIC 4-digit level MFN tariffs are computed as simple averages
over the HS 6 digit tariffs.
Applied MFN tariffs vary substantially both across sectors within countries and across countries for a given sector. For example, U.S. manufacturing tariffs in 2004 averaged 2.4 percent,
with a minimum of zero and a maximum of 350 percent. As an example of cross-country variation, for a sector like SIC 3631 (Household Cooking Equipment), applied MFN tariffs varied
between zero and 29 percent, with an average of 3.15 percent.
Our analysis focuses on tariffs on final goods in the domestic market. In some regressions,
we also control for the tariffs applied to imported inputs, using the variable Input Tariff k,c . This
is a weighted average of 4-digit SIC applied MFN tariffs, using normalized IO-coefficients from
the US input-output table as weights. To proxy for the level of protection faced by exporters in
foreign markets, we use the variable Export Tariff k,c . We construct this variable by weighting
tariffs in destination markets with bilateral sectoral export shares using information from the
UN Comtrade database.
The variable MFN share k,c measures the fraction of imports to which MFN tariffs apply, for
each country and sector. This excludes imports from countries with which the importer has a
preferential trade agreement, which do not face tariff restrictions. The higher is this share, the
more sensitive its domestic prices should be to MFN tariffs. For example, the U.S. will have low
MFN shares in sectors in which it imports a lot from its NAFTA trading partners (Canada and
Mexico). In these sectors, the MFN tariff that the U.S. imposes on other WTO members will
have little impact on domestic prices. In contrast, the effect may be substantial in sectors where
most imports originate in countries with which the U.S. has no preferential trade agreement.
To distinguish between firms selling only in the domestic market and exporting firms, we
30
If information on applied MFN tariffs is unavailable for that year, we use the closest available data point in a
five year window around 2004 (2002-2006), with priority given to earlier years. For example, if data are available
for 2003 and 2005, but not 2004, the 2003 data are chosen.
15
construct two measures. The dummy variable Domestic f is constructed from WorldBase and
takes the value of 1 if firm f does not report to be an exporter. The variable Import-competing k,c
is a country-sector specific measure of import-competition constructed using information from
Comtrade. This is a dummy indicating whether a firm operates in one of the 25 percent most
import-competing sectors, based on the ratio of a country’s total imports/exports by sector.
We also collect information on all regional trade agreements in force in 2004 from the WTO
Regional Trade Agreements Information System (RTA-IS).31 We construct the dummy RTAc,c0
that equals one when countries c and c0 belong to a regional trade agreement formed under
Article XXIV of the GATT.32 To distinguish between different types of RTAs, we construct the
dummy variables Customs Union c,c0 and Free Trade Area c,c0 . We expect the former, which imply
a common external tariff and no internal trade barriers, to have a stronger effect on organizational
convergence than the latter, which permit member countries to maintain different external tariffs.
4.5
Other controls
We collect a number of country- and sector-specific variables to control for alternative factors
emphasized in the literature on vertical integration.
In terms of country-specific variables, the empirical and theoretical literatures have studied
the role of institutional characteristics and financial development.33 We use the variable Legal
Quality c to proxy for the quality of a country’s institutions. This is the variable “rule of law”
from Kaufmann, Kraay, and Mastruzzi (2003), which is a weighted average of a number of
variables (perception of incidences of crime, effectiveness and predictability of the judiciary, and
enforceability of contracts) between 1997 and 1998. The variable ranges from 0 to 1 and is
increasing in the quality of institutions. The variable Financial Development c measures private
credit by deposit money banks and other financial institutions as a fraction of GDP for 2004
and is taken from Beck, Demigurc-Kunt, and Levine (2006).
We also construct the variable Capital Intensity k , using data from the NBER-CES manufacturing industry database (Bartelsmann and Gray, 2000) at the 4-digit-SIC level. In line with
the literature, capital intensity is defined as the log of total capital expenditure relative to value
added averaged over the period 1993-1997.
To control for domestic industry concentration, we construct Herfindahl k,c indices using in31
Available online (http://rtais.wto.org/UI/PublicMaintainRTAHome.aspx).
This variable does not include a number of preferential trade agreements under the Enabling Clause that do
not imply the full elimination of trade barriers.
33
Poor legal institutions may affect vertical integration decisions through their impact on the severity of holdup problems. Financial development may affect integration positively if a sufficient level may be necessary for
upstream and downstream firms to be able to integrate, or negatively insofar as integration facilitates borrowing
and therefore substitutes for poor financial institutions. As Acemoglu, Johnson and Mitton (2009) note, the
effect of each of these variables may be ambiguous when considered separately and there may be more robust
predictions of their combined effect.
32
16
formation on sales of all plants in a given country and sector.34
To proxy for the degree of product differentiation, we use two dummy variables. The variable Homogeneous1 k is equal to 1 when a sector is homogeneous according to the well-known
classification by Rauch (1999).35 The dummy variable Homogeneous2 k,c is constructed using
information on sector-country-specific import demand elasticities estimated by Broda, Greenfield and Weinstein (2006).36 It takes value 1 whenever the elasticity is above the median for
the country. Broda, Greenfield and Weinstein (2006) show that sectors with more homogeneous
products are characterized by higher import demand elasticities.
In some specifications, we include the variable Size f , using information on firm-level employment from WorldBase. Since firm size is clearly endogenous to vertical integration, we always use
predicted size as an instrument, constructed by regressing firm size on sector-country dummies.
Similarly, we construct labor productivity measured as firm sales divided by employment. Again,
we instrument this variable using predicted (with sector-country dummies) labor productivity.
In the regressions on organizational convergence, we also use a number of bilateral variables
from CEPII: bilateral Distance measured as the simple distance between the most populated
cities (in km), dummies for Contiguity c,c0 , for Common Language c,c0 (official or primary), and
Colonial Relationship c,c0 (current or past). In some specifications, we also include the variable
Difference GDP c,c0 for the year 2004 constructed from the World Development indicators.
5
Tariffs and vertical integration
In this section, we assess the empirical validity of the main prediction of our theoretical model
that higher prices for the final good lead to more vertical integration at the firm level. To
examine the organizational effects of trade policy, we exploit variation in applied MFN output
tariffs across countries and sectors (the following section exploits time-series variation in the
degree of protection faced by firms). We estimate the following reduced form regression model:
Vf,k,c = α + β1 Tariffk,c + β2 Xf,k,c + δk + δc + f,k,c ,
(4)
where Xf,k,c is the vector of explanatory variables, δk and δc are sector and country dummies
and f,k,c is an error term with E(f,k,c |Xf,k,c , δk , δc ) = 0. Thus, the effect of Tariff k,c on Vf,k,c is
causal conditional on covariates.
We study the determinants of Vf,k,c , the vertical integration index of firm f , with primary
sector k, located in country c, as defined in (3). Since the distribution of vertical integration
34
These include sales by foreign-owned plants that operate in the given country-sector.
Rauch (1999) classifies products according to three different types: homogeneous goods, which are traded in
organized exchanges; goods that are are not traded in organized exchanges, but for which a published reference
price can be found; and differentiated goods, which fall under neither of the two previous categories.
36
We thank David Weinstein for making these data available to us.
35
17
indices is rather skewed (see Figure 1), we use log of one plus Vf,k,c as our dependent variable.37
Our main regressor of interest is the variable Tariff k,c , which is the log of one plus the MFN
tariff applied to output in sector k by country c.38 Our model predicts that higher final good
tariffs within an industry should lead firms in that industry to be more vertically integrated. We
thus expect the coefficient β1 to be positive.39
The vector Xf,k,c includes a series of firm- and sector-country-specific controls, that we will
discuss below. We also include sector fixed effects at the 4-digit SIC level (δk ), which allows us to
capture cross-industry differences in technological or other determinants of vertical integration
(e.g. a sector’s capital intensity). Finally, we add country fixed effects (δc ), which capture
cross-country differences in institutional determinants of vertical integration (e.g. a country’s
level of financial development and the quality of its contracting institutions) and also control
for country-specific differences in the way firms are sampled. Given that tariffs vary only at the
sector-country level, while the dependent variable varies at the firm level, we cluster standard
errors at the sector-country level.
Table 1 reports the results of estimations in which we test the main predictions of our
theoretical framework. Column (1) presents the results of the basic specification, which includes
only the variable Tariff and country and sector fixed effects. The estimated coefficient for the
tariff is 0.02 (implying a tariff elasticity of vertical integration of the same magnitude) and
strongly significant. Consistent with the first prediction of our theoretical model, higher tariffs
lead firms to be more vertically integrated. To the extent that domestic prices do not fully
adjust to tariff changes, our estimates can be interpreted as a lower bound on the impact of
prices on vertical integration. As discussed below, for competitive sectors, in which firms cannot
strategically adjust their prices in response to tariff changes, our estimates imply a price elasticity
of vertical integration of up to 2.1.
In columns (2) and (3) we verify whether the effect of domestic tariffs on organization is
larger for firms that operate only in the domestic market (for which only this price should
affect the degree of vertical integration). To do so, we interact the variable Tariff k,c with
two dummy variables: Domestic f , which is constructed using information on from WorldBase
and takes the value of 1 if firm f does not report to be an exporter; and Import-competing k,c ,
which is constructed using information from Comtrade and indicates whether a firm operates
37
We have also used the log of the vertical integration index (removing zero observations), obtaining similar
results. There are very few zeros in the dependent variable, so there is no need to perform a Tobit analysis. All
results not shown due to space considerations are available upon request.
38
Tariffs are expressed in ad-valorem terms. In the main specifications, we use log of (one plus the tariff) in
order to be able to include zero tariffs. Although the distribution of tariffs is extremely skewed, the log of (one
plus the tariff) is approximately normally distributed. In all regressions we scale vertical integration indices and
tariffs so that adding unity to the variables is quantitatively irrelevant. This implies that the coefficients can
still be interpreted as tariff elasticities of vertical integration. In alternative specifications, we used log vertical
integration and log tariffs, obtaining very similar results.
39
We have also performed a series of estimations including a quadratic term for Tariff k,c , finding no evidence
of a non-monotonic relationship between tariffs and vertical integration.
18
in one of the 25 percent most import-competing sectors, based on the ratio of a country’s total
imports/exports by sector. We expect the coefficients on the interaction terms to be positive.
In column (2) the coefficient for tariffs (which measures the impact of tariffs on vertical
integration for exporters) is positive but insignificantly different from zero. On the other hand,
the coefficient on the interaction term is positive, strongly significant and similar in magnitude
to the baseline specification. Thus, import tariffs have a significant effect on vertical integration
only for firms that sell exclusively in the domestic market. In column (3), we use the alternative
measure to identify firms that do not export to foreign markets. Again, the coefficient on the
interaction term is positive and significant at the five-percent level, indicating that import tariffs
have a bigger impact on vertical integration decisions for firms that operate in import-competing
sectors.
In column (4) we verify whether tariffs have a larger impact on vertical integration when
the share of imports to which they apply is larger (implying a bigger effect on domestic prices).
To do this, we include the variable MFN share k,c , capturing the fraction of imports to which
MFN tariffs apply in a given country and sector, as well as the interaction between this variable
and the tariff. The coefficient in the first row now measures the impact of MFN tariffs when no
imports are subject to them (i.e. in a sector in which a country imports only from regional trading
partners). Not surprisingly, this coefficient is not significant, since in this case MFN tariffs should
have no impact on domestic prices. The interaction term is instead positive and significant at
the one-percent level, indicating that the effect of MFN tariffs on vertical integration is positive
and increasing in their importance for import volumes.
In columns (5)-(8) we repeat the same specifications, adding interaction terms that have been
emphasized in previous studies on vertical integration. In particular, Acemoglu, Johnson and
Mitton (2009) find evidence that contracting costs and financial development have a stronger impact on vertical integration in more capital-intensive sectors. We thus introduce two interaction
terms: one between Capital Intensity k and Financial Development c and the other one between
Capital Intensity k and Legal Quality c . The coefficient on the first interaction term is positive
and significant, indicating that more capital intensive sectors are more integrated in countries
with more developed financial markets. The second interaction term has the expected negative
sign but it is not significant. In all specifications, our results on the effect of tariffs on firm-level
vertical integration are unaffected.
5.1
Alternative mechanisms
Our theoretical analysis focuses on a perfectly competitive setting, in which firms are price
takers. According to our model, tariff changes should affect organizational choices through their
impact on product prices: higher tariffs should raise prices and thus increase the incentives for
vertical integration.
19
In reality, tariff changes may also affect vertical integration decisions through their impact
on the degree of competition faced by firms. In particular, Aghion, Griffith and Howitt (2006)
suggest a U-shaped relationship between competition and vertical integration: a small increase in
competition reduces a producer’s incentive to integrate by improving the outside options of nonintegrated suppliers and hence raising their incentive to make relationship-specific investments;
too much competition raises the producer’s incentive to integrate, by allowing non-integrated
suppliers to capture most of the surplus.
To isolate the organizational effects of product prices, in Table 2 we restrict our analysis to
highly competitive sectors, in which tariffs changes should have little or no effect on the degree
of competition. In all specifications, we impose two restrictions to define competitive industries: i) there are at least 20 domestic firms operating in that country and sector; ii) goods are
homogeneous. Further restrictions are imposed in some specifications, as discussed below. To
distinguish between differentiated and homogeneous sectors, we adopt two alternative methodologies: in Panel A, we use the dummy variable Homogeneous1 k , which identifies industries in
which goods are traded in organized exchanges, classified as homogeneous according to Rauch
(1999); in Panel B, we use instead the variable Homogeneous2 k,c , which identifies sectors with
high import demand elasticities according to Broda, Greenfield and Weinstein (2006). Notice
that the sample is much larger in the bottom panel, since the variable Homogeneous2 k,c varies
at the country-sector level.
In the baseline specifications of columns (1)-(2), competitive sectors are identified based
only on the two criteria discussed above. Additional restrictions are imposed in the rest of the
table. Columns (3)-(4) include only sectors with low levels of protection (Tariff k,c < 10%), in
which domestic firms face a high level of foreign competition. In columns (5)-(6) the sample
is restricted to sectors in which some foreign-owned firms have plants in the domestic market,
further increasing the competitive pressure on domestic firms. In columns (7)-(8) we exclude
concentrated sectors, i.e. industries for which the Herfindahl k,c index is above 0.1.
In all specifications, the coefficient for Tariff k,c is positive and significant at least at the fivepercent level. The results of Table 2 allow us to identify the price-level effects of tariff changes
on firm boundaries, abstracting from possible competition effects. In line with our theoretical
model, these results suggest that higher import tariffs lead domestic firms to be more vertically
integrated, by increasing the price at which they sell their final products.
Table 2 shows that, in competitive sectors, the tariff elasticity of vertical integration ranges
between 0.025 and 0.1.40 Given that tariffs are expressed in ad-valorem terms rather than in
units, at the average tariff rate of around 5 percent, these estimates imply that the price elasticity
of vertical integration is much greater: the estimates of the upper panel of Table 2 imply that
40
In terms of standard deviations, a one standard deviation increase in tariffs implies an up to 0.12 standard
deviation increase in vertical integration.
20
the price elasticity of vertical integration ranges between 0.61 (column 1) and 2.1 (column 8).41
Price changes can thus have significant effects on firm boundaries.
The fact that tariffs continue to positively affect vertical integration when restricting the
analysis to homogeneous and highly competitive sectors is consistent with the predictions of our
theoretical model. By contrast, it is difficult to reconcile this result with models in which tariffs
affect firm boundaries through their impact on the degree of competition.
Another possible explanation for our results could be that protected firms have more disposable cash to acquire their suppliers. Notice that this explanation relies on the fact that firms
are credit constrained, so that the amount of cash available matters for takeovers decisions. If
this is the reason behind the positive impact of tariffs on vertical integration, we would expect
the effect to be stronger in sectors and countries in which credit constraints are more severe.
To verify this, we have interacted the tariff variable with the inverse of our measure Financial
Developmentc and with a standard measure of financial dependence, which capture the extent
of financial market imperfections in different countries and sectors.42 We have tried different
specifications (e.g. including the interaction terms separately or together, including a triple interaction between tariffs, financial dependence and the inverse of financial development). In all
cases, the interaction terms were insignificant and the sign and significance of the tariff coefficient
was unaffected, suggesting that cash availability is not the reason behind the positive effect of
tariffs on vertical integration.43
5.2
Omitted variables
Our analysis shows that firms are more vertically integrated when import tariffs on their final
product are higher. In this section, we deal with endogeneity concerns, to establish a causal
relationship between tariffs and organization decisions. As argued above, reverse causality is
unlikely to be a problem in our analysis, since there is no reason to believe that firms’ ownership
structures affect applied MFN tariff rates. However, MFN tariffs on final products could be
correlated with omitted variables that may also affect vertical integration decisions.
In what follows, we show that our results are robust to controlling for two sets of potential
41
To see this, denote the domestic price of good k in country c as pk,c = (1+tk,c )Pk , where Pk is the world price
∂pk,c tk,c
tk,c
∂Pk tk,c
of good k and tk,c is the tariff. Then the tariff elasticity of domestic prices is given by ∂tk,c
pk,c = 1+tk,c + ∂tk,c Pk ,
where the first part on the right is the direct impact of an ad-valorem tariff on domestic prices (< 1) and the
∂Vk,c tk,c
second term is the terms of trade effect (< 0). Define ∂tk,c
Vk,c ≡ β. The price elasticity of vertical integration can
h
i−1
∂V
pk,c
tk,c
∂Pk tk,c
then be written as ∂pf,k,c
= β 1+t
+ ∂t
. Abstracting from terms of trade effects, and substituting
k,c Vk,c
k,c
k,c Pk
∂V
p
k,c
the average tariff of 5 percent, we obtain ∂pf,k,c
= β ∗ 21.
k,c Vk,c
42
Following Rajan and Zingales (1998), we define a firm’s external dependence on finance as capital expenditures
minus cash flow from operations divided by capital expenditures. Our sectoral measure of external dependence on
finance is the median firm’s external dependance on finance in a given sector in the U.S in the period 1999-2006,
constructed from Compustat.
43
The results of these regressions are omitted due to space considerations, but are available upon request.
21
omitted variables. First, we include measures of input tariffs and export tariffs, which are
correlated with output tariffs and may also affect vertical integration decisions.44 Second, we
control for firm size, labor productivity and industry concentration, which can affect the degree
of protection through their impact on lobbying pressure (e.g. Mitra, 1999; Bombardini, 2008)
and may also be correlated with firms’ ownership structures.
The results of these regressions are presented in Table 3. For comparison, in the first two
columns, we report the results of the baseline specifications. In column (3)-(8), we add additional
controls, first one by one and then simultaneously: Input Tariff k,c , Export Tariff k,c , Herfindahl k,c ,
Size f (instrumented with predicted size) and Labor Productivity f (instrumented with predicted
labor productivity). Notice that, in all specifications, the coefficient on Tariff is positive, highly
significant, and very stable. These results indicate that omitted variables are not a concern and
that higher output tariffs lead to more vertical integration.
Of the additional controls, firm size and labor productivity have a positive and significant
effect on organization, a result that continues to hold in all additional robustness checks. The fact
that more productive firms are more likely to be integrated can be explained by the theoretical
model, in which firms with higher levels of exogenous productivity have more incentives to
vertically integrate at any given price.45
The estimated coefficient on input tariffs in Table 3 is positive and significant, though this
result disappears in some of the robustness checks (e.g. Table A-4). It should be stressed that
our theory generates no clearcut predictions concerning the effects of input tariffs on boundary
choices: whether higher input prices strengthen or weaken the incentives for integration depends
on whether inputs sales are part of the revenue of the enterprise (in which case higher input
tariffs will lead to higher integration) or part of its (contractible) costs. For example, if an
automobile manufacturer produces more stampings than it needs and sells the remainder on the
market, then the sale of stampings will enter its revenue; in this case, the higher the price of
stampings, the higher the incentives to integrate. On the other hand, if stampings are purchased
on the open market, an increase in their price will diminish revenue and reduce the incentives
to integrate.
Finally, notice that the coefficient on Export Tariff k,c is insignificant in all specifications and
in the additional robustness checks (see Table A-4). This finding can easily be explained by our
model: if countries are small, tariffs in foreign markets should have no impact on domestic prices
and thus on vertical integration decisions by domestic firms.46
44
The simple correlation of output tariffs with input tariffs is 0.78 and the one with export tariffs is 0.31.
While we cannot formally test whether the instruments (average firm size and labor productivity in a given
country-sector pair) are valid, because the model is exactly identified, we are confident that including predicted
firm size or productivity as controls does not introduce any endogeneity bias, since the coefficient of the variable
Tariff is unaffected when including these controls.
46
By contrast, this result is hard to reconcile with models of vertical integration choices by multinationals,
in which tariffs affect location decisions, and these are inextricably intertwined with boundary decisions. For
example, in the model by Diez (2010) tariffs in one country (North or South) should always decrease vertical
45
22
5.3
Additional robustness checks
In line with the predictions of our theoretical model, our empirical analysis shows that higher
output tariffs lead domestic firms to be more vertically integrated. This effect is stronger for
firms serving only the domestic market — for which organizational choices should depend solely
on domestic prices — and operating in sectors in which a smaller share of imports originate from
regional trading partners — for which MFN tariffs should have a larger impact on domestic
prices.
The results presented in Tables 1-3 already show that the organizational effects of tariffs
are robust to the inclusion of many different controls that account for alternative drivers of
vertical integration decisions. In a series of additional robustness checks, we have verified that
higher tariff on final goods continue to have a positive and significant effect on firm-level vertical
integration when using different econometric methodologies or focusing on alternative samples.
In Table A-4 in the Appendix, we have reproduced all the specifications of Table 3, using
a Poisson quasi-maximum likelihood (PQML) estimator to assess the effect of tariffs on vertical integration. The rationale for this exercise is that Santos Silva and Tenreyro (2006) have
shown that for log-linear models the OLS estimator gives biased estimates in the presence of heteroscedasticity and have suggested the PQML estimator as an alternative with good statistical
properties. Vertical integration is now estimated in levels, which allows to include observations
for which the dependent variable is zero, while the explanatory variables are in logs and can thus
be interpreted as elasticities. Standard errors are again clustered at the sector-country level.
Our main result on the impact of output tariffs is unaffected: in all specifications, the coefficient
for the output tariff is always positive and significant. Input and export tariffs have instead no
significant effect on firm-level vertical integration.
The organizational effects of output tariffs were also unaffected in a series of additional
estimations discussed below. The results of these specifications are omitted from the paper due
to space considerations, but are available upon request.
We have restricted the sample to OECD countries. Our methodology for constructing vertical
integration indices better applies to these countries: since they are more similar to the United
States in terms of technology, it is less problematic to use U.S. input-output matrix to measure
technological linkages between sectors.47
We have used an alternative measure of vertical integration, constructed based on all the
P j
, where Nf is the number
firm’s activities rather than its primary activity: V f,k,c = N1f j Vf,k,c
of industries in which firm f is active. The coefficients for MFN tariffs remained strongly
significant but, not surprisingly, they dropped slightly in magnitude.
integration in the other country, so tariffs faced by exporters should have a negative effect on vertical integration.
47
Moreover, most MFN tariffs applied by OECD countries in 2004 coincide with the bindings set in the Uruguay
Round of multilateral trade negotiations (1986-1994), particularly in non-agricultural sectors (Bchir, Jean and
Laborde, 2006). Governments have thus no room to adjust them under the pressure of import-competing firms.
23
In our analysis, we cluster standard errors at the sector-country level. Alternatively, we have
tried clustering at the sector level, at the country level and two-way clustering at the sector and
country level. In all cases, the coefficient for Tariff k,c remained strongly statistically significant.
We have also carried out the analysis on two additional samples of firms. First, we have
restricted the sample to countries for which we observe at least 1000 plants of sufficient size in
order to eliminate any bias that may arise from differences in sampling across countries. Second,
we have included multinational firms to the main sample. As noted above, since multinationals
have plants in different countries, it is hard to identify with precision the tariffs that affect their
organization decisions; we have thus split them into separate firms by country and used the
primary activity of the respective domestic ultimate to identify the relevant tariff. For each of
these samples, we have reproduced all the specifications of Table 3, adding a dummy variable
for multinational status for the sample including multinationals. As expected, the coefficient for
Tariff k,c remained always positive and strongly significant.
6
Time-series evidence: China’s accession to the WTO
As noted in the introduction, China’s accession to the WTO in 2001 is arguably the only major
trade liberalization episode that has occurred in the last decade, for which we can use D&B data
to construct vertical integration measures. To be accepted as a member of the WTO, China
agreed to undertake a series of important commitments to better integrate in the world economy
and offer a more predictable environment for trade and foreign investment in accordance with
WTO rules.48 In particular, China had to substantially expand market access to goods from
foreign countries, reducing its import tariffs from an average of 13.3 percent in 2001 to 6.8
percent by the end of the implementation period.49
Our identification strategy is based on the comparison of two periods, a pre-accession one
and a post-accession one, to verify whether firm-level vertical integration was reduced by more in
those sectors that experienced larger tariff cuts. We thus construct vertical integration measures
for all Chinese manufacturing firms that are in the WorldBase dataset for the years 1999 (pre
accession) and 2007 (post accession), following the same procedure described in Section 4.3.
We use 2007 instead of 2004 as the post-accession period because we expect firms’ ownership
structure to react slowly to price changes induced by tariff reductions.
Figure 2 provides the histograms of the MFN tariffs applied by China in 1999 and 2007.
The sample is based on those manufacturing sectors for which we observe firms (with at least
20 employees, excluding multinationals) in both years, consisting of almost 29,000 firms that we
observe in at least one year. For the sectors in this sample, applied tariffs fell from an average 20
48
A detailed list of China’s commitments can be found in its Protocol of Accession. China’s accession implied
few trade policy changes for other WTO members, since most of them had already been granting it MFN status.
49
The implementation period lasted until 2010, though most tariff reductions had to be completed by 2005.
24
to an average of 9.9 percent between 1999 and 2007, with a lot variation across sectors.50 At the
same time, the average level of vertical integration for the sample of firms declined from 0.111
to 0.084.51 Figure 3 visualizes the leftward shift in the distribution of VI indices between 1999
and 2007.
Figure 2: Chinese import tariffs, 1999 and 2007
In what follows, we examine whether Chinese firms have adjusted their vertical integration
structure following WTO accession in response to the tariff reductions. To this purpose, we run
two sets of regressions. First, we use a very similar specification as in our main test (4), using
only those sectors for which we observe some firms in both 1999 and in 2007:
Vf,k,t = α + β1 Tariffk,t + β2 Xf,k,t + δk + δt + f,k,t .
(5)
Here, Xf,k,t is again a vector of controls, which includes Public f,t , a dummy for public ownership
from Worldbase, and Herfindahl k,t . We control for Public f,t since public ownership is very common in China and may be correlated with vertical integration. Again, we expect the coefficient
of Tariff k,t to be positive. Notice that, by controlling for sector fixed effects, we exploit the time
variation of tariffs within sectors. Specifically, the tariff coefficient is identified by the deviation
of firm-level vertical integration from its sector mean that is due to the time variation in tariffs
50
The maximum reduction in tariffs was 415 percent (SIC 3578, Calculating and Accounting Machines), the
median reduction was 51 percent. Only in a few sectors, tariffs did not change or actually increased (e.g. SIC
2084 Wines, Brandy and Brandy Spirits).
51
One may be concerned that tariff levels and reductions may be endogenous to industry characteristics, for
example because industries with larger firms, more concentrated industries, or industries with more prevalence
of public ownership would lobby for higher initial tariff levels and smaller subsequent tariff reductions. If on the
other hand, these sectors are systematically different in terms of vertical integration, one may spuriously obtain
negative correlations between vertical integration and tariffs. In our sample, however, this is not the case: the
level of tariffs in 1999 is neither significantly correlated with sector-level average firm size, nor with industry
concentration or public ownership in the same year. Moreover, changes in tariffs between 1999 and 2007 are also
not significantly correlated with the level of the previous variables in 1999.
25
Figure 3: Chinese vertical integration indices, 1999 and 2007
relative to their sector mean. Given that we only consider sectors for which we can observe firms
in both periods, sector averages of vertical integration and tariffs are well identified. General
trends in vertical integration, which may be due to other reforms that occurred in China over
the sample period, are picked up by time dummies.52
In a second set of regressions, we focus on within-firm variation in VI indices. Unfortunately,
the overlap between the firms sampled in 1999 and 2007 is small. Once we exclude multinationals
and plants with less than 20 employees, as we have done in our earlier analysis, there are 145
firms that we can observe in both years. For this set of firms, we take time differences of equation
(5) and estimate
∆Vf,k = α + β1 ∆ Tariffk + β2 ∆Xf,k + ∆f,k .
(6)
Again, we expect the coefficient of ∆Tariff k to be positive. In these regressions, we control
not only for changes in firm size, industry concentration and public ownership status, but alternatively also for changes in the degree of state ownership by sector, by including the variable
Privatization. This measures the fraction of government-owned firms that were privatized in a
given sector (at the 2-digit industry level) between 1999 and 2004 and is taken from Bai, Lu and
Tao (2009).
Table 4 presents the results for both sets of regressions. Columns (1)-(4) reports the results
for the regressions with sector dummies. In all specifications, we find a positive and significant
(at the one percent level) coefficient on the tariff variable, implying larger reductions in vertical
integration in sectors that have experienced larger tariff reductions. The coefficient magnitude
is around 0.03, which is slightly larger than our cross-section estimates. The coefficient of the
public-ownership dummy is positive and highly significant, indicating that publicly owned firms
52
In these regressions, unobserved firm-specific effects are assumed to be common for all firms in a given sector.
26
are more vertically integrated. Finally, the level of industry concentration has no significant
effect on vertical integration.
Turning to the specification in differences, in columns (5)-(10) we obtain similar results.
The coefficient of tariff changes is always positive, significant and similar in magnitude to the
specification with sector dummies. In column (6) we add changes in industry concentration as
control, which leaves the tariff coefficient unaffected. Column (7) adds change in public ownership
status as control, which is insignificant and does not change the tariff coefficient. In column (8)
we alternatively use Privatization as a control, which is again insignificant and also leaves the
coefficient of tariffs unchanged. Finally, in columns (9) and (10), we simultaneously control
for changes in public ownership structure and changes in industry concentration by adding the
Herfindahl index. While changes in tariffs remain positive and significant, changes in industry
concentration and changes in public ownership have no significant effects.
7
Trade policy and organizational convergence
The purpose of this section is to assess the validity of the predictions of our model on how
trade policy affects organizational convergence between countries, through its impact on price
convergence.
To measure organizational convergence, we construct sector-country-specific measures of vertical integration by regressing firm-level vertical integration on industry-sector dummies and
firm size. The estimate for the sector-country dummy gives us a measure of the average level of
vertical integration of industry k in country c, denoted by V̂k,c .
We first examine whether cross-country differences in sectoral organizational structure are affected by differences in tariffs. For a given country-pair c, c0 , we expect organizational differences
to be smaller in sectors characterized by similar levels of protection.
To verify this, we estimate the following model:
|V̂k,c − V̂k,c0 | = α + β1 | Tariffk,c − Tariffk,c0 | + β2 |Xk,c − Xk,c0 | + δk + δc,c0 + k,c,c0 .
(7)
The dependent variable is the absolute difference between countries c and c0 in the estimated
vertical integration indices for sector k. All differences are expressed in logs. The main regressor
of interest is the (log of the) absolute difference between these countries’ MFN tariffs in sector
k. The term |Xk,c − Xk,c0 | captures differences in other sector-country characteristics that may
affect the degree of organizational convergence. Note that, because we are including dyad fixed
effects (δc,c0 ), β1 is identified by the cross-sectoral variation in the tariff difference for a given
country pair. To allow for correlation of the errors between sectors for a given country pair, we
cluster standard errors by dyad.
The results of these regressions are reported in Panel A of Table 5. In column (1), the only
27
explanatory variable is the log-difference in MFN tariffs. In line with our predictions, we find
that, for a given country-pair differences in sectoral vertical integration indices are significantly
(at the one percent level) larger in sectors in which differences in MFN tariffs are larger. A 100
percent increase in the difference in MFN tariffs leads to a roughly 0.9 percent increase in the
difference in vertical integration indices. The second column adds interactions between Capital
Intensity and differences in Financial Development and Legal Quality. The coefficient on the
difference in MFN tariffs remains relatively unchanged in magnitude and is significant at the 5
percent level.
We next examine the relation between the degree of sectoral organizational convergence and
common membership in a regional trade agreement. In contrast to the regressions on tariff
differences, a causal interpretation of these regression results is more difficult, since it is possible
that countries that are generally more similar are more likely to form RTAs.
To assess the validity of our final empirical prediction, we explore how RTAs affect the extent
to which two countries have similar vertical integration structures at the industry level.
|V̂k,c − V̂k,c0 | = α + β1 RT Ac,c0 + β2 Xc,c0 + δk + δc + δc0 + k,c,c0 .
(8)
The dependent variable is as in model (7), expressed as before in logs. The main regressor of
interest is now RTAc,c0 , a dummy that equals one if countries c and c0 are members of the same
RTA. The vector Xc,c0 captures a series of bilateral controls, such as dummies for contiguity, common language, and colonial relationship, as well as variables that capture the distance between
countries, and differences in GDP (differences expressed in logs of absolute values). Finally,
we include sector fixed effects (δk ) and country fixed effects (δc and δc0 ). Standard errors are
clustered by country-pair.
The results of these regressions are reported in Panel B of Table 5. In column (1), in which
we include only a dummy for regional trade agreements, the coefficient of RTA is negative and
significant at the one-percent level. This implies that the difference in vertical integration indices
for a country pair in an RTA is about 9.2 percent smaller than for a country pair without an
RTA. The results for an alternative specification, which separates customs unions (CUs) from
free trade areas (FTAs), are presented in column (2). As expected, the quantitative impact on
organizational convergence is greater for CUs than for FTAs. Country pairs that belong to the
same CU have a approximately 18.5% smaller difference in organizational structure than country
pairs without a RTA, while membership to FTAs has no significant impact on differences in
organizational structure. In column (3), we keep the coefficients for CUs and FTAs separate and
add a series of bilateral control variables that may have an impact on similarity of organizational
structure. The coefficient for CUs is reduced somewhat in size, but remains significant at the
10 percent level. Contiguity and common language have a significant negative effect on the
difference in vertical integration indices, while differences in GDP have a significant positive
28
effect. Colonial relationship and distance do not affect the degree of organizational convergence.
As done for the results presented in Section 5, we have verified that our findings on trade policy and organizational convergence are robust to using different samples of firms (e.g. restricting
the sample of countries included in the analysis, including multinationals). In all specifications,
are results continue to hold: tariff differences have a significant positive effect on differences
in vertical integration; and membership in RTAs, and CUs in particular, continues to reduce
differences in vertical integration among member countries.
8
Conclusions
Recent work in organization economics suggests that product prices can have a key impact on
integration decisions. This mechanism arises in a setting in which vertical integration increases
productive efficiency, but does so at a cost. At low prices, the productivity gains from integration
have little value, so non-integration is preferred; at higher prices, integration becomes worthwhile.
We assess the empirical validity of this mechanism by studying the organizational effects of
trade policy, which provides a source of price variation that is exogenous to firms’ ownership
decisions. We have constructed firm-level vertical integration indices for a large set of countries
and industries and exploited cross-country and cross-sectoral differences in applied MFN tariffs,
as well as time variation in tariffs resulting from China’s accession to the WTO, to study the
impact of prices on firm boundaries. In line with our model’s predictions, we find that higher
tariffs on final goods lead firms to be more vertically integrated; this effect is stronger for nonexporting firms, which are more sensitive to domestic prices, and for sectors in which domestic
prices are more sensitive to import tariffs.
Positive correlations between prices and vertical integration have been observed in many
industries. For example, a 1989 report on the beer industry by the British Monopolies and
Mergers Commission found that retail prices were higher in integrated than non-integrated pubs.
Similarly, Hastings (2004) noted that increases in gasoline prices in California in the 1990’s were
associated with increases in the number of vertically integrated gasoline stations. Policymakers
have often drawn a causal inference from this correlation: vertical integration causes higher
prices, though industrial economists have been rather more reticent to draw such a conclusion.53
Though it is still certainly possible that vertical integration may raise prices in some industries in
the manner suggested by the foreclosure theories, our analysis suggests that a positive correlation
between vertical integration and prices may also reflect causality working in the opposite way:
higher prices may induce more vertical integration.
53
Recent empirical studies have questioned this inference in the context of specific industries, either by providing
alternative explanations for a positive correlation between prices and integration (Hastings, 2004) or by showing
that the correlation is actually negative (Hortaçsu and Syverson, 2007).
29
References
Acemoglu, D., P. Aghion, R. Griffith, and F. Zilibotti (2010). “Vertical Integration and Technology: Theory and Evidence,” Journal of European Economic Association, 8, 1-45.
Acemoglu, D., S. Johnson, and T. Mitton (2009). “Determinants of Vertical Integration: Financial Development and Contracting Costs,” Journal of Finance 63, 1251-1290.
Aghion, P., R. Griffith, and P. Howitt (2006). “Vertical Integration and Competition,” American
Economic Review Papers and Proceedings 96, 97-102.
Alfaro, L., and A. Charlton (2009). “Intra-Industry Foreign Direct Investment,” American
Economic Review 99, 2096-2119.
Antras, P. (2003). “Firms, Contracts, and Trade Structure,” Quarterly Journal of Economics
118, 1375-1418.
Antras, P., and R. W. Staiger (2012). “Offshoring and the Role of Trade Agreements,” American
Economic Review 102, 3140-3183.
Bai, C., J. Lu, and Z. Tao (2009). “How Does Privatization Work in China,” Journal of
Comparative Economics 37, 453-470.
Baker, G., and T. Hubbard (2003). “Make Versus Buy in Trucking: Asset Ownership, Job
Design and Information,” American Economic Review 93, 551-572.
(2004). “Contractibility and Asset Ownership: On-Board Computers and Governance in
US Trucking,” Quarterly Journal of Economics 119, 1443-1479.
Bartelsman, E. J., and W. Gray (1996). “The NBER Manufacturing Productivity Database,”
NBER Technical Working Paper 205.
Beck T., A. Demigurc-Kunt, and R. Levine (2006). A New Database on Financial Development
and Structure (1960-2006).
Bchir, H., S. Jean, and D. Laborde (2006). “Binding Overhang and Tariff-Cutting Formulas,”
Review of World Economics 142, 207-232.
Bernard, A. B., S. J. Redding, and P. K. Schott (2011). “Multi-Product Firms and Trade
Liberalization,” Quarterly Journal of Economics, 126, 1271-1318.
Black, S. E., and P. E. Strahan (2002). “Entrepreneurship and Bank Credit Availability,”
Journal of Finance 57, 2807-2833.
Bloom, N., and J. Van Reenen (2007). “Measuring and Explaining Management Practices
Across Firms and Countries,” Quarterly Journal of Economics 122, 1351-1408.
Bloom, N., R. Sadun, and J. Van Reenen (2010). “Does Product Market Competition Lead
Firms to Decentralize?,” American Economic Review 100, 434-438.
Bloom, N., R. Sadun, and J. Van Reenen (2012). “The Organization of Firms Across Countries,”
forthcoming, Quarterly Journal of Economics.
Bolton, P. and M. D. Whinston (1993). “Incomplete Contracts, Vertical Integration, and Supply
Assurance,” Review of Economic Studies 60, 121-148.
30
Bombardini, M. (2008). “Firm Heterogeneity and Lobby Participation,” Journal of International
Economics 75, 329-348.
Bradford, S. (2003). “Paying the Price: Final Goods Protection in OECD Countries,” Review
of Economics and Statistics 85,24-37.
Broda, C., J. Greenfield, and D. Weinstein (2006). “From Groundnuts to Globalization: A
Structural Estimate of Trade and Growth,” NBER Working Paper No. 12512.
Cadot, O., J. de Melo, and M. Olarreaga (1999). “Regional Integration and Lobbying for Tariffs
against Nonmembers,” International Economic Review 40, 635-657.
Caves, R. E. (1975). “Diversification, Foreign Investment and Scale in North American Manufacturing Industries,” Canadian Public Policy 2, 274-276.
Coase, R. H. (1937). “The Nature of the Firm,”Economica 4, 386-405.
Conconi, P., Legros, P., and A. F. Newman (2012). “Trade Liberalization and Organizational
Change,” Journal of International Economics 86, 197-208.
Diez, F. J. (2010). “The Asymmetric Effects of Tariffs on Offshoring Industries: How North/South
Tariffs Affect Intra-Firm Trade.” Federal Reserve Bank of Boston working paper 10-4.
Eckel, C., and P. J. Neary (2010). “?Multi-product Firms and Flexible Manufacturing in the
Global Economy,?” Review of Economic Studies 77, 188-217.
Fan, J. P. H., and L. H. Lang (2000). “The Measurement of Relatedness: An Application to
Corporate Diversification,” Journal of Business 73, 629-660.
Finger, J. M., H. K. Hall, and D. R. Nelson (1982). “The Political Economy of Administered
Protection,” American Economic Review 72, 452-466.
Forbes, S. J., and M. Lederman (2009). “Adaptation and Vertical Integration in the Airline
Industry,” American Economic Review 99, 1831-1849.
(2011). “Does Vertical Integration Affect Firm Performance? Evidence from the Airline
Industry,” Rand Journal of Economics 41, 765-790.
Grossman S. J., and O. D. Hart (1986). “The Costs and Benefits of Ownership: A Theory of
Vertical and Lateral Integration,” Journal of Political Economy 94, 691-719.
Grossman, G. M., and E. Helpman (1994). “Protection for Sale,”American Economic Review
84, 833-850.
Grossman, G. M., and E. Helpman (2002). “Integration Versus Outsourcing In Industry Equilibrium,”Quarterly Journal of Economics 117, 85-120.
Guadalupe, M., and J. Wulf (2012). “The Flattening Firm and Product Market Competition:
The Effect of Trade Liberalization,” American Economic Journal: Applied Economics 4,
105-127.
Harrison, A. E., I. Love, and M. S. McMillian (2004). “Global Capital Flows and Financing
Constraints,” Journal of Development Economics 75, 269-301.
31
Hart, O. D., and B. R. Holmström, (2010). “A Theory of Firm Scope,” Quarterly Journal of
Economics, 125, 483-513.
Hart, O. D., and J. H. Moore (1990). “Property Rights and the Nature of the Firm,” Journal
of Political Economy 98, 1119-1158.
Hart, O. D., and J. Tirole (1990). “Vertical Integration and Market Foreclosure,” Brookings
Papers on Economic Activity, 205-276.
Hastings, J. (2004). “Vertical Relationships and Competition in Retail Gasoline Markets: Empirical Evidence from Contract Changes in Southern California,” American Economic Review 94, 317-328.
Holmström, B. and P. Milgrom (1991). “Multitask Principal-Agent Analyses: Incentive Contracts, Asset Ownership, and Job Design,” Journal of Law, Economics and Organization
7, 24-52.
Hortaçsu, A., and C. Syverson (2007). “Cementing Relationships: Vertical Integration, Foreclosure, Productivity and Prices,” Journal of Political Economy 115, 250-301.
Hortaçsu, A., and C. Syverson (2009). “Why do Firms Own Production Chains? ”mimeo.
Hummels, D., J. Ishii, and K. Yi (2001). “The Nature and Growth of Vertical Specialization in
World Trade,” Journal of International Economics 54, 75-96.
Johnson, J. P. and D. P. Myatt (2003). “Multiproduct Quality Competition: Fighting Brands
and Product Line Pruning,” American Economic Review 93, 748-774.
Joskow, P. (1987). “Contract Duration and Relationship-Specific Investments: Empirical Evidence from Coal Markets,” American Economic Review 77, 168-185.
Kaufmann, D., A. Kraay, and M. Mastruzzi (2004). “Governance Matters III: Governance
Indicators for 1996, 1998, 2000, and 2002,” World Bank Economic Review 18, 253-287.
Klapper, L., L. Laeven, and R. Rajan (2006). “Entry Regulation as a Barrier to Entrepreneurship,” Journal of Financial Economics 82, 591-629.
Klein, B., R. G. Crawford, and A. Alchian (1978).“Vertical Integration, Appropriable Rents and
the Competitive Contracting Process,” Journal of Law and Economics, 21, 297-326.
Lafontaine, F., and M. Slade (2007). “Vertical Integration and Firm Boundaries: The Evidence”
Journal of Economic Literature, 45, 629-685.
Legros, P., and A. F. Newman (2008). “Competing for Ownership,” Journal of the European
Economic Association 6, 1279-1308.
Legros, P., and A. F. Newman (2013). “A Price Theory of Vertical and Lateral Integration,
forthcoming, Quarterly Journal of Economics.
McLaren, J. (2000). “Globalization and Vertical Structure,” American Economic Review 90,
1239-1254.
Melitz, M. J. (2003). “The Impact of Trade on Intra-Industry Reallocations and Aggregate
Industry Productivity,” Econometrica 71, 1695-1725.
32
Mitra, D. (1999). “Endogenous Lobby Formation and Endogenous Protection: A Long-Run
Model of Trade Policy Determination,” American Economic Review 89, 1116-1134.
Nunn, N. (2007). “Relationship-Specificity, Incomplete Contracts, and the Pattern of Trade,”
Quarterly Journal of Economics 122, 569-600.
Ornelas, E., and J. Turner (2008). “Trade Liberalization, Outsourcing, and the Hold Up Problem,” Journal of International Economics 74, 225-241.
Ordover, J. A., G. Saloner, and S. C. Salop (1990). “Equilibrium Vertical Foreclosure,” American
Economic Review 80, 127-142.
Perry, M. K. (1989). “Vertical Integration: Determinants and Effects,” in R. Schmalensee and
R. Willig (eds.) Handbook of Industrial Organization I, North-Holland.
Rajan, R., and L. Zingales (1998). “Financial Dependence and Growth,” American Economic
Review 88, 559-586.
Ramondo, N., V. Rappoport, and K. Ruh (2012). “Horizontal versus Vertical FDI: Revisiting
Evidence about U.S. Multinationals,” mimeo.
Rauch, J. (1999). “Networks Versus Markets in International Trade,” Journal of International
Economics 48, 7-35.
Riordan, M. H. (2008). “Competitive Effects of Vertical Integration,” in P. Buccirossi (ed.)Handbook
of Antitrust Economics, MIT Press.
Salinger, M. A. (1988). “Vertical Mergers and Vertical Foreclosure,” Quarterly Journal of
Economics 103, 345-356.
Santos-Silva J. M. C and Silvana Tenreyro (2006). “The Log of Gravity,” Review of Economics
and Statistics 88, 641-658.
Stuckey, J. A. (1983). Vertical Integration and Joint Ventures in the Aluminium Industry.
Harvard University Press, Cambridge, MA.
Topkis, D. (1998). Supermodularity and Complementarity, Princeton University Press.
Vives, X. (2000). Oligopoly Pricing: Old Ideas and New tools, MIT Press.
Williamson, O. (1971). “The vertical integration of production: market failure considerations,”
American Economic Review 61, 112123.
(1975). Markets and Hierarchies: Analysis and Antitrust Implications, Free Press.
Woodruff, C. (2002). “Non-contractible Investments and Vertical Integration in the Mexican
Footwear Industry, International,” Journal of Industrial Organization 20, 1197-1224.
33
Table 1: Tariffs and vertical integration
Tariffk,c
(1)
0.0203***
(0.0061)
Domesticf
Tariffk,c
x Domesticf
(2)
0.0034
(0.0088)
-0.0926***
(0.0108)
0.0214***
(0.0054)
Import-competingk,c
(3)
0.0160**
(0.0063)
(4)
0.0042
(0.0080)
(5)
0.0202***
(0.0600)
(6)
0.0035
(0.0086)
-0.0923***
(0.0109)
0.0212***
(0.0054)
-0.0362***
(0.0135)
0.0155**
(0.0070)
Tariffk,c
x Import-competingk,c
MFN sharek,c
(7)
0.1600***
(0.0061)
-0.0351***
(0.0134)
0.0148**
(0.0070)
34
-0.0237
(0.0190)
0.0245***
(0.0074)
Tariffk,c
x MFN sharek,c
(8)
0.005
(0.0081)
-0.0215
(0.0190)
0.0230***
(0.0075)
Capital Intensityk
x Financial Developmentc
0.0322**
(0.0142)
0.0321**
(0.0144)
0.0297**
(0.0148)
0.0293**
(0.0145)
Capital Intensityk
x Legal Qualityc
-0.0833
(0.0564)
-0.0823
(0.0573)
-0.0774
(0.0589)
-0.0735
(0.0572)
196,586
386
0.117
YES
196,586
386
0.119
YES
196,586
386
0.117
YES
196,586
386
0.117
YES
# Observations
# Sectors
R2
Sector and Country Fixed Effects
196,586
386
0.117
YES
196,586
386
0.119
YES
196,586
386
0.117
YES
196,586
386
0.117
YES
Notes: Robust standard errors clustered at the sector-country level in parentheses denoting *** 1%, **5%, and *10% significance. The sample
includes firms ≥ 20 employees in the manufacturing sector, excluding multinationals. Dependent variable: log of one plus Vf,k,c , the vertical
integration index of firm f , with primary sector k, located in country c. Explanatory variables are in logs, except MFN tariffs, where we
use log of one plus the tariff. The variable Tariffk,c is the MFN tariff imposed by country c in sector k. The dummy variables Domesticf
and Import-competingk,c are alternative ways to capture firms that only sell in the domestic market. Domesticf identifies firms that do not
export. Import-competingk,c identifies the top 25 percent most import-competing sectors, as measured by a country’s total imports/exports.
MFN sharek,c measures the fraction of imports of good k by country c that are subject to the MFN tariff, i.e. do not originate from countries
with which country c has a regional trade agreement. Capital Intensity k is the total capital expenditures divided by value added. Financial
Development c measures private credit by deposit money banks and other financial institutions as a fraction of GDP. The variable Legal Quality c
proxies for the quality of a country’s institutions
Table 2: Tariffs and vertical integration, competitive industries
Tariffk,c
Capital Intensityk
x Financial Developmentc
(1)
(2)
Homogeneous sectors,
many firms
0.0290**
0.0292**
(0.0117)
(0.0117)
0.0063
(0.0524)
Capital Intensityk
x Legal Qualityc
Observations
# Sectors
R2
Sector and Country Fixed Effects
35
Tariffk,c
Capital Intensityk
x Financial Developmentc
-0.0615
(0.1950)
13,095
56
0.073
YES
13,095
56
0.073
YES
Panel B:
(1)
(2)
Homogeneous sectors,
many firms
0.0257*** 0.0248***
(0.0083)
(0.0084)
0.0216
(0.0248)
Capital Intensityk
x Legal Qualityc
Observations
# Sectors
R2
Sector and Country Fixed Effects
Panel A: Homogeneous sectors based on Rauch (1999)
(3)
(4)
(5)
(6)
(7)
(8)
Homogeneous sectors,
Homogeneous sectors,
Homogeneous sectors,
many firms, low tariffs many firms, foreign presence many firms, low concentration
0.0380**
0.0381**
0.0316**
0.0315**
0.0747***
0.0982***
(0.0149)
(0.0154)
(0.0145)
(0.0144)
(0.0228)
(0.0216)
0.0203
- 0.0035
0.68
(0.0436)
(0.4420)
(0.1610)
-0.2330
(0.1440)
78,437
337
0.107
YES
1.027
(0.7290)
11,279
11,279
10,918
10,918
8,539
8,539
54
54
53
53
37
37
0.052
0.052
0.068
0.068
0.047
0.047
YES
YES
YES
YES
YES
YES
Homogeneous sectors based on Broda, Greenfield and Weinstein (2006)
(3)
(4)
(5)
(6)
(7)
(8)
Homogeneous sectors,
Homogeneous sectors,
Homogeneous sectors,
many firms, low tariffs many firms, foreign presence many firms, low concentration
0.0345*** 0.0341*** 0.0363***
0.0363***
0.0648***
0.0639***
(0.0085)
(0.0084)
(0.0101)
(0.0102)
(0.0112)
(0.0116)
0.0041
0.0560
-0.0267
(0.0245)
(0.0367)
(0.0299)
-0.1240
(0.0875)
78,437
337
0.106
YES
0.0499
(0.1870)
-0.0952
(0.1170)
69,823
328
0.099
YES
69,823
328
0.099
YES
- 0.3018**
(0.1306)
69,980
309
0.111
YES
69,980
309
0.111
YES
-0.0840
(0.1400)
50,315
234
0.087
YES
50,315
234
0.087
YES
Notes: Robust standard errors clustered at the sector-country level are in parentheses; denoting *** 1%, **5%, and *10% significance. The sample
includes firms ≥ 20 employment in the manufacturing sector, excluding multinationals. In all columns, the sample is restricted to industries in
which at least 20 domestic firms operate in a given country, and in which products are homogeneous. In Panel A, homogeneous sectors are defined
using the variable Homogeneous1k based on Rauch (1999); in Panel B, we use instead the variable Homogeneous2 k,c based on Broda, Greenfield
and Weinstein (2006). In columns (3)-(8), we impose further restrictions: columns (3)-(4) include only sectors in which Tariffk,c < 10%; columns
(5)-(6) include only sectors in which some foreign firms operate in the domestic market; in columns (7)-(8), the sample is restricted to sectors in
which Herfindahlk,c ≤ 0.1. Dependent variable: log of one plus Vf,k,c , the vertical integration index of firm f , with primary sector k, located in
country c. Explanatory variables are in logs, with the exception of MFN tariffs, where we use log of one plus the tariff. Tariffk,c is the MFN tariff
imposed by country c in sector k. Capital Intensity k is the total capital expenditures divided by value added. Financial Development c measures
private credit by deposit money banks and other financial institutions as a fraction of GDP. The variable Legal Quality c proxies for the quality of a
country’s institutions.
Table 3: Tariffs and vertical integration, controlling for possible omitted variables
Tariffk,c
(1)
0.0203***
(0.0061)
Capital Intensityk
x Financial Developmentc
Capital Intensityk
x Legal Qualityc
(2)
0.0202***
(0.0060)
0.0322**
(0.0142)
(3)
0.0203***
(0.0074)
0.0333**
(0.0161)
(4)
0.0190***
(0.0058)
0.0387**
(0.0154)
(5)
0.0214***
(0.0065)
0.0425**
(0.0176)
(6)
0.0210***
(0.0059)
0.0401**
(0.0187)
(7)
0.0212***
(0.0059)
0.0319**
(0.0146)
(8)
0.0198***
(0.0057)
0.0390**
(0.0189)
(9)
0.0183***
(0.0061)
0.0507**
(0.0234)
-0.0833
(0.0564)
-0.0785
(0.0617)
-0.100*
(0.0604)
-0.106
(0.0682)
-0.103
(0.0711)
-0.0836
(0.0577)
-0.0944
( 0.0705)
-0.127
(0.0885)
0.00252
(0.0052)
0.0330**
(0.0141)
-0.0038
(0.0060)
0.0267***
(0.0054)
0.0464***
( 0.0170)
-0.0049
(0.0063)
0.0141
( 0.0316)
0.0440***
(0.0111)
0.0287***
(0.0069)
178,448
386
0.125
YES
133,545
311
0.132
YES
Input Tariffk,c
0.0391***
(0.0142)
Export Tariffk,c
Herfindahlk,c
0.0128
(0.0228)
Sizef
36
0.0352***
(0.0075)
Labor Productivityf
# Observations
# Sectors
R2
Sector and Country Fixed Effects
196,586
386
0.117
YES
196,586
386
0.117
YES
154,915
311
0.119
YES
185,630
386
0.123
YES
146,228
311
0.125
YES
178,199
386
0.117
YES
196,586
386
0.122
YES
Notes: Robust standard errors clustered at the sector-country level are in parentheses denoting *** 1%, **5%, and *10% significance. The sample
includes firms ≥ 20 employment in the manufacturing sector, excluding multinationals. Dependent variable: log of one plus Vf,k,c , the vertical
integration index of firm f , with primary sector k, located in country c. Explanatory variables are in logs, with the exception of output tariffs, input
tariff and export tariff, where we use log of one plus the tariff. Tariffk,c is the MFN tariff imposed by country c in sector k. Capital Intensity k
is the total capital expenditures divided by value added. Financial Development c measures private credit by deposit money banks and other
financial institutions as a fraction of GDP. The variable Legal Quality c proxies for the quality of a country’s institutions. Import Tariffk,c is the
tariff imposed by country c on inputs of good k. Export Tariffk,c is the tariff faced by firms exporting good k from country c. Herfindahlk,c is the
Herfindahl index of sector k in country c. In columns (7)-(9) we report 2SLS estimates. Sizef measures firm size (instrumented with employment
predicted with sector-country dummies). Labor productivityf is measured as sales per worker (instrumented with sales per worker predicted with
sector-country dummies).
Table 4: Tariffs and vertical integration, China’s accession to the WTO
Tariffk,t
(1)
0.0313***
(0.0096)
Herfindahlk,t
(2)
0.0356**
(0.0160)
-0.0040
(0.0113)
Publicf,t
(3)
0.0313***
(0.0096)
0.0055***
(0.0010)
(4)
0.0357**
(0.0160)
-0.0040
(0.0113)
0.0059***
(0.0015)
Change Tariffk
(5)
(6)
(7)
(8)
(9)
(10)
0.0364*
(0.0202)
0.0459*
(0.0234)
-0.0371
(0.0292)
0.0364*
(0.0203)
0.0364*
(0.0205)
0.0456*
(0.0234)
-0.0307
(0.0297)
-0.0351
(0.0320)
0.0451*
(0.0230)
-0.0406
(0.0301)
Change Herfindahlk
Change Publicf
0.0007
(0.0135)
Privatizationk
# Observations
# Sectors
R2
Sector Fixed Effects
Time Fixed Effects
-0.0351
(0.0798)
28,928
88
0.921
YES
YES
13,682
88
0.902
YES
YES
28,928
88
0.921
YES
YES
13,682
88
0.902
YES
YES
145
88
0.042
NO
YES
75
75
0.087
NO
YES
145
88
0.042
NO
YES
145
88
0.042
NO
YES
0.0574
(0.1180)
75
75
0.104
NO
YES
75
75
0.089
NO
YES
Notes: Robust standard errors clustered at the sector level in parentheses denoting *** 1%, **5%, and *10% significance. In columns
(1)- (4), the dependent variable is the log of one plus Vf,k , the vertical integration index of firm f , with primary sector k; in
columns (5)-(10), it is the change in the log of one plus the vertical integration index between 1999 (pre accession) and 2007 (post
accession). The sample includes Chinese firms observed in 1999 and/or 2007 with ≥ 20 employees in the manufacturing sector,
excluding multinationals. Tariffk is the log of (one plus) MFN tariff applied by China in sector k. Publicf is a firm-level dummy for
public ownership, Privatizationk is a sector-level measure of privatization and Herfindahlk is the Herfindahl index for sector k.
Table 5: Trade policy and organizational convergence
Panel A: Tariff differences
(1)
0.0089***
(0.0034)
Capital Intensityk
x Difference Financial Developmentc,c0
(2)
0.0086**
(0.0037)
0.0020
(0.0066)
Capital Intensityk
x Difference Legal Qualityk,c,c0
0.0419***
(0.0062)
Difference Tariffsk,c,c0
# Observations
212,770
171,908
# Country pairs
80
80
R2
0.164
0.164
Sector and Country-pair Fixed Effects
YES
YES
Panel B: Regional Trade Agreements
(1)
-0.0921***
(0.0235)
RTAc,c0
Customs Unionc,c0
Free Trade Areac,c0
(2)
(3)
-0.185***
(0.0376)
-0.0404
(0.0266)
-0.0760*
(0.046)
0.0203
(0.0264)
Distancec,c0
0.0188
(0.0146)
-0.196***
(0.0754)
-0.119***
(0.0313)
0.0663
(0.0421)
0.0389***
(0.0087)
Contiguityc,c0
Common Languagec,c0
Colonial Relationshipc,c0
Difference GDPc,c0
# Observations
# Country pairs
R2
Sector and Country Fixed Effects
299,649
101
0.109
YES
299,649
101
0.109
YES
299,649
101
0.111
YES
Notes: Robust standard errors clustered at the country-pair level in parentheses denoting *** 1%, **5%, and *10% significance. Dependent variable:
log of (one plus) the absolute difference between countries c and c0 in the
estimated vertical integration index of firms with primary sector k. The
variable Difference Tariffsk,c,c0 is the difference between the MFN tariff imposed by country c and c0 in sector k. Capital Intensity k is the total capital
expenditures divided by value added. Difference Financial Development cc0
measures differences in private credit by deposit money banks and other financial institutions as a fraction of GDP. The variable Difference in Legal
Quality c proxies differences in the quality of institutions across between two
countries. The variables Contiguityc,c0 , Colonial Relationshipc,c0 , Common
Languagec,c0 and Difference GDPc,c0 capture bilateral geographical, cultural
and economic linkages.
38
Appendix
Table A-1: Main sample
WB code
ALB
ARG
AUS
AUT
BEL
BEN
BFA
BGD
BGR
BOL
BRA
CAN
CHE
CHL
COL
CRI
CZE
DEU
DNK
ECU
ESP
FIN
FRA
GAB
GBR
GHA
GRC
GTM
HND
HUN
IDN
IND
IRL
ISR
ITA
JAM
JOR
JPN
KEN
KOR
Freq.
4
998
5,079
1,464
928
4
8
6
360
55
5,594
7,469
1,150
454
550
176
1,736
19,302
425
183
2,322
448
8,965
3
6,622
81
2,231
93
77
2,346
233
2,592
587
1,538
8,426
43
148
34,441
134
3,060
Percent
Cum.
WB code
0.00
0.51
2.58
0.74
0.47
0.00
0.00
0.00
0.18
0.03
2.85
3.80
0.58
0.23
0.28
0.09
0.88
9.82
0.22
0.09
1.18
0.23
4.56
0.00
3.37
0.04
1.13
0.05
0.04
1.19
0.12
1.32
0.30
0.78
4.29
0.02
0.08
17.52
0.07
1.56
0.00
0.51
3.09
3.84
4.31
4.31
4.32
4.32
4.50
4.53
7.38
11.18
11.76
11.99
12.27
12.36
13.24
23.06
23.28
23.37
24.55
24.78
29.34
29.34
32.71
32.75
33.89
33.93
33.97
35.17
35.29
36.60
36.90
37.68
41.97
41.99
42.07
59.59
59.66
61.21
MAR
MDG
MEX
MLI
MOZ
MUS
MWI
MYS
NER
NIC
NLD
NOR
NZL
OMN
PAK
PER
PHL
PNG
POL
PRT
PRY
ROM
RWA
SAU
SEN
SGP
SLV
SWE
TGO
THA
TTO
TUN
TUR
TZA
UGA
URY
USA
VEN
ZAF
ZMB
Total
Freq.
Percent
Cum.
603
18
2,641
13
16
46
2
3,101
1
21
676
847
959
67
4
888
351
4
446
5,433
50
614
2
314
47
790
129
689
4
507
79
991
2,557
24
37
114
52,917
231
1
17
196,586
0.31
0.01
1.34
0.01
0.01
0.02
0.00
1.58
0.00
0.01
0.34
0.43
0.49
0.03
0.00
0.45
0.18
0.00
0.23
2.76
0.03
0.31
0.00
0.16
0.02
0.40
0.07
0.35
0.00
0.26
0.04
0.50
1.30
0.01
0.02
0.06
26.92
0.12
0.00
0.01
100.00
61.52
61.53
62.87
62.88
62.89
62.91
62.91
64.49
64.49
64.50
64.84
65.27
65.76
65.80
65.80
66.25
66.43
66.43
66.66
69.42
69.45
69.76
69.76
69.92
69.94
70.35
70.41
70.76
70.76
71.02
71.06
71.57
72.87
72.88
72.90
72.96
99.87
99.99
99.99
100.00
Notes: The sample includes all firms in the 2004 WorldBase dataset by Dun & Bradstreet, which are located in
WTO member countries and have primary activities in manufacturing sectors. It excludes firms with less than
20 employees and multinationals.
39
Table A-2: Summary statistics
Sample
Plants
Connected plants
Multi-plant firms
Single-plant firms
Firms
Country variables
Vertical Integration Indexf
Sizef
Labor productivityf
Domesticf
Import-competingk,c
Tariffk,c
Input Tariffk,c
Export Tariffk,c
MFN Sharek,c
Herfindahlk,c
Homogeneous1k
Homogeneous2k,c
Capital Intensityk
Financial Developmentc
Legal Qualityc
Country-pair variables
Difference Ver. Int. Indexk,c,c0
Regional Trade Agreementc,c0
Free Trade Agreementc,c0
Customs Unionc,c0
Distancec,c0
Contiguityc,c0
Colonial Relationshipc,c0
Common Languagec,c0
Difference GDPc,c0
Median
0.044
38.000
11.506
0
0.702
2.480
2.546
5.654
0.564
0.053
0
0
-2.857
0.332
0.545
Median
-1.593
0.000
0.000
0.000
9.017
0.000
0.000
0.000
0.450
Mean
0.063
98.936
11.446
0.233
0.705
4.849
3.994
6.611
0.545
0.132
0.081
0.491
-2.902
0.554
0.583
Mean
-1.707
0.263
0.148
0.115
8.629
0.041
0.020
0.122
0.201
Std. Dev.
0.063
472.395
1.082
0.423
0.571
7.253
4.954
5.039
0.351
0.188
0.273
0.499
0.458
0.479
0.209
Std. Dev.
1.614
0.440
0.355
0.319
0.965
0.139
0.178
0.328
1.812
N
225,212
29,214
6,830
189,756
196,586
N
196,586
196,586
178,448
196,586
196,586
196,586
154,915
185,630
196,586
178,199
196,586
173,587
387
80
80
N
299,649
299,649
299,649
299,649
299,649
299,649
299,649
299,649
299,649
Sources: Vertical Integration Indexf , Size f , Labor Productivity f , Domestic f and Herfindahl k,c constructed using
plant-level data from 2004 WorldBase, Dun & Bradstreet. Sample includes manufacturing firms with at least 20
employees and excludes multinationals. Tariff k,c from the World Integrated Trade Solution (WITS); Input Tariff k,c , Export Tariff k,c , and MFN Share k,c constructed using data from WITS and the UN Comtrade database.
Import-Competing k,c constructed using data from Comtrade. Information on regional trade agreements from
the WTO. Homogeneous1 k from Rauch (1999), Homogeneous2 k,c constructed using data from Broda, Greenfield
and Weinstein (2006). Capital Intensity k from NBER-CES manufacturing industry database. Financial Development c from Beck, Demigurc-Kunt and Levine (2006). Legal Quality c from Kaufmann, Kraay, and Mastruzzi
(2004). GDP c from the World Bank. Contiguity c,c0 , Colonial Relationship c,c0 , and Common Language c,c0 from
CEPII. Vertical Integration Index f ,Tariff k,c , Input Tariff k,c , Export Tariff k,c , Size f , Herfindahl k,c , and MFN
Share k,c are in levels; all other variables (with the exception of indicator variables) are in logs.
40
Table A-3: Vertical integration by 2-digit SIC industry
Industry
TEXTILES
APPAREL
CHEMICALS
PRIMARY METAL PRODUCTS
ELECTRICAL MACHINERY
TRANSPORTATION EQUIPMENT
PETROLEUM REFINING
LEATHER
RUBBER AND PLASTICS
MACHINERY, EXCEPT ELECTRICAL
MANUFACTURING NEC
LUMBER AND WOOD PRODUCTS
FOOD AND KINDRED PRODUCTS
TOBACCO MANUFACTURES
STONE, CLAY, GLASS, & CONCRETE
FABRICATED METAL PRODUCTS
PRINTING AND PUBLISHING
SCIENTIFIC INSTRUMENTS
PAPER AND ALLIED PRODUCTS
FURNITURE AND FIXTURES
SIC
22
23
28
33
36
37
29
31
30
35
39
24
20
21
32
34
27
38
26
25
VI index
0.115
0.111
0.098
0.091
0.089
0.067
0.062
0.062
0.060
0.060
0.059
0.059
0.056
0.053
0.049
0.039
0.039
0.036
0.034
0.022
Notes: Data from 2004 WorldBase data, Dun & Bradstreet. Sample includes firms ≥ 20 employment in the
manufacturing sector, excluding multinationals.
41
Table A-4: Tariffs and vertical integration, Poisson quasi-maximum likelihood estimator
Capital Intensityk
x Financial Developmentc
(2)
0.0121**
(0.0058)
0.0385**
(0.0165)
(3)
0.0281***
(0.0086)
0.0545**
(0.0267)
(4)
0.0114*
(0.0061)
0.0443**
(0.0178)
(5)
0.0258***
(0.0092)
0.0651**
(0.0293)
(6)
0.0117*
(0.0064)
0.0484**
(0.0208)
(7)
0.0126**
(0.0058)
0.0399**
(0.0165)
(8)
0.0108*
(0.0064)
0.0469**
(0.0205)
(9)
0.0191*
(0.0105)
0.0822**
(0.0361)
Capital Intensityk
x Legal Qualityc
-0.153**
(0.0629)
-0.232**
(0.1030)
-0.164**
(0.0679)
-0.265**
(0.1140)
-0.189**
(0.0744)
-0.163***
(0.0629)
-0.176***
(0.0738)
-0.337**
(0.1370)
0.0043
(0.0050)
-0.0042
(0.0208)
0.0002
(0.0070)
0.0322***
(0.0079)
0.0210
(0.0210)
-0.0011
(0.0072)
0.0284
(0.0520)
0.0565***
(0.0142)
0.0365***
( 0.0105)
133,607
386
YES
99,017
311
YES
Tariffk,c
(1)
0.0126**
(0.0058)
Input Tariffk,c
-0.0122
(0.0201)
Export Tariffk,c
Herfindahlk,c
0.0027
(0.0290)
Sizef
42
0.0438***
(0.0079)
Labor Productivityf
# Observations
# Sectors
Sector and Country Fixed Effects
149,992
386
YES
149,992
386
YES
116,086
311
YES
142,364
386
YES
110,448
311
YES
133,413
386
YES
149,992
386
YES
Notes: Robust standard errors clustered at the country-sector level are in parentheses denoting *** 1%, **5%, and *10% significance. The
sample includes firms ≥ 20 employment in manufacturing sectors, excluding multinationals. Dependent variable: Vf,k,c , the vertical integration
index of firm f , with primary sector k, located in country c. Explanatory variables are in logs and can be interpreted as elasticities. Tariffk,c
is the log of the MFN tariff imposed by country c in sector k. Capital Intensity k is the total capital expenditures divided by value added.
Financial Development c measures private credit by deposit money banks and other financial institutions as a fraction of GDP. The variable
Legal Quality c proxies for the quality of a country’s institutions. Import Tariffk,c is the tariff imposed by country c on inputs of good k.
Export Tariffk,c is the tariff faced by firms exporting good k from country c. Herfindahlk,c is the Herfindahl index of sector k in country c.
Sizef is predicted (with sector-country dummies) firm size, Labor Productivityf is predicted (with sector-country dummies) sales per worker.